Colorado, University of California Publications, Colorado Delta, material, Natural Landscape, Salton Sink, California, Pattie Basin, dunes, Salton Basin, Volcano Lake, Peninsular Range, Santa Clara Mesa, Punta San Felipe, Sierra Juarez, vegetation, Cerro Prieto, stream flow, University of California, steep slope, Yuma, The Salton Sea, desert ranges, alluvium, desert range, beach line, river terraces, Gulf of California, The Natural Landscape
LOWER CALIFORNIA STUDIES IV. THE NATURAL LANDSCAPE OF THE COLORADO DELTA BY FRED B. KNIFFEN INTRODUCTION The delta of the Colorado River lies within the area included between the parallels of 31° and 33° 31' N, and between the meridians 114° 30 and 116° W (fig. 1). It includes that portion of this area in which alluvium derived from the waters of the Colorado forms the dominant surface soil constituent. Fig. 1 Sketch map showing location of the Colorado Delta.
150 University of California Publications in Geography [VOI... 5 Climatically this area is a desert, various portions of it being differently named by local usuage. It is a part of the great desert area continued in the east as the Sonora Desert and to the north as the Colorado Desert. The well watered portion of the delta has never been a desert in the. popular mind, and quite possibly this distinction is a legitimate one. Physiographically it falls into the Basin Range Province, a region of isolated diastrophic blocks and intervening valleys, which extends northward to the Columbia plateaus and southward far into Mexico. This province is bounded sharply on the west by the Peninsular Range, visible from our area and even touching it. The delta is included within the Lower Sonoran division of the Lower Austral life-zone.' But again, both as to flora and fauna the delta stands in striking contrast to the truly desert area about it. It is the steaming jungle which breaks abruptly into the sparsely vegetated aridity which surrounds it. Description of the area-- The delta of the Colorado, as here interpreted, extends from the Salton Sea on the north to Punta San Felipe on the south; from the Peninsular Range or its outliers on the west to Santa Clara Mesa on the east. The maximum length of the area is approximately one hundred fifty miles, its greatest width about forty miles. To the west lies the great crystalline mass of the Peninsular Range. South df the international boundary the eastern front of this range presents a remarkably regular scarp wall. In the area north of the border it forms a series of salients and reentrants which result in great step-like irregularities. The Santa Rosa 1VIounthins form one of the salients ; accompanying it are the corresponding reentrants, Palm Canyon and Borrego Valley. To the east a mesa rises sharply above the delta and extends many miles in a great flat-lying plain before encountering the desert ranges. The mesa, evidently of marine origin, is composed of recent sands and gravels. The latter frequently exhibit excellent developments of pedre gales or desert pavements. Nowhere are there clear evidences of stream dissection or even of stream channels; only along the edge of the mesa where it stands at a height of from thirty to fifty feet above the delta alluvium is there a suggestion of erosion. Striking relief features of this generally smooth plain are the two areas of 1 C. HART MERRIAM, LIFE ZONES AND CROP ZONES OF THE UNITED STATES. U. S. DEPT. OF AGR., BIOLOGICAL SURVEY, BULL. 10:41 (1898).
1932] Kniffen: The Natural Landscape of the Colorado Delta 151 dunes; the one, the Algodones Sand Hills, the other, the sand hills lying about twenty-five miles east of the mesa edge in Sonora. Within the area bounded by the Peninsular Range and the Santa Clara Mesa and to some extent interrupting the continuity of the alluvial surface, are a number of desert ranges and isolated peaks. Occupying a central position with respect to the long axis of the delta, and closely paralleling the scarp front of the Peninsular Mountains, lies a continuous series of mountain structures to the whole of which the name Sierra de los Cucopas is commonly applied. This system is separated from the Sierra Juarez, as the Peninsular Range is here called, by the width of the alluvium-filled basin of Laguna Salada. In the major part the Cucopa system is composed of granite, but there are areas of extrusives and metamorphics. The two rock types stand in marked contrast in the forms which they reveal in weathering. Several of the constituent parts of this system, in their steepness of slope to the east and more gentle slope to the west, suggest a structural origin similar to that of the Sierra Juarez where the same slope pattern is repeated on a larger scale. The great, bare, alluvial plain which extends south beyond the tip of the Cucopas to Punta San Felipe is bounded on the west by an irregular series of volcanoes and mountains of interbedded tuff and lava to which the name Sierra de las Pintas is applied. These mountains are extremely irregular in form as well as in extent, are frequently made up of strikingly colored rocks, and are practically destitute of vegetation. Between Las Pintas and Sierra Juarez lies another desert range, somewhat more distinctly a single system than the former, which is called Sierra de las Tinajas. Crystalline rocks are present in this area but they are subordinate in significance to a great mass of lavas and tuffs. A portion of the northernmost part of this range is a basalt-capped mesa, the basalt extending with interruptions to the wall of Sierra Juarez. Another semi-isolated mountain mass is Superstition Mountain, which lies on the western side of Imperial Valley. This has a core composed of crystalline rocks which are flanked by Tertiary sandstones interbedded with basalt and tuff. This dry, barren mountain has an extreme outlier in an exposure of sandstone found a few miles west of the town of Imperial, one of the few rock exposures within the delta proper.
152 University of California Publications in Geography [ VOL. 5 Most aptly termed by Sykes the "three corner stones of the delta" are the conspicuous, isolated masses of Cerro Prieto, Pilot Knob, and Punta San Felipe.2 Pilot Knob stands at the contact of mesa and delta and marks the emergence of the Colorado from its restricted channel onto the broad alluvial plain that is its delta. The mountain is composed of granite, cut by later volcanics. The extinct volcano, Cerro Prieto or Black Butte, stands conspicuously alone, several miles removed from the eastern flanks of the Cucopas. It also marks the contact of the mesa sands, which lie between it and the mountains to the west, with the river alluvium. The rounded granitic dome of Punta San Felipe stands on the western shore of the Gulf of California and marks the southern limit of surface exposures of river alluvium. South of it the beaches are sandy ; to the north the alluvial plain widens rapidly as the head of the gulf narrows into the river mouth. The general configuration and slopes of the delta show an adaptation to the unique conditions under which the structure was formed. Instead of a free open sea, the river has debouched into a closely confined gulf. As a result, strong tidal action has influenced the growth of the delta. It is generally true that in areas of strong tidal action great deltas do not form, so that, as might be expected, the Colorado Delta exhibits some abnormal characteristics. What may be termed the delta crest follows roughly a line from Pilot Knob to Cerro Prieto. Its western terminus is marked by the dry bed of Volcano Lake. South of the crest the slope its quite uniform to the gulf, with a gradient of less than two feet to the mile. North of the crest the slope to the Salton Sea is much greater (fig. 2). Subordinate and secondary to the above are the slopes of the alluvium-surfaced Pattie Basin in which lies Laguna Salada. This basin is a true bolson, having no drainage outlet. To the north it is separated from the Salton area by a rocky ridge which connects the northern tip of the Cucopas, Centinela, with the Sierra Juarez. The channel of ingress for the silt-bearing waters has been through the opening which lies between the southern point of the Cucopa system, Sierra Mayor, and the desert ranges to the south. The lowest point in this gap has an elevation of about ten feet above the sea. The lowest point in the basin, in the extreme northeastern corner, lies several feet below sea level
. The lower margin is along the eastern 2 Godfrey Sykes, The Delta and Estuary of the Colorado River, Geog. Rev., 16:235 (1926).
1932] Kniff en: The Natural Landscape of the Colorado Delta 153 side of the basin ; to the west is a long slope partly formed by sands and gravels derived from the Sierra Juarez. In very few instances is there any difficulty in delimiting sharply the contact of river alluvium with soils of other origin. In many places it is the easy distinction between fine silt and coarse sand or gravel. Throughout most of the area the depositional limits are marked by relief forms. The term "mesa" as used in this area means flat-topped sand and gravel surfaces which lie marginal to the alluvial plain and present to it a steep-walled ascent of from thirty to fifty feet. Wherever the line of contact lies some distance from the mountains the mesa is clearly distinguishable. Where the river alluvium approaches the mountains the contact is equally clear but of a somewhat different nature. The eastern face of the Cucopas is in large part bordered by an apron of colluvial material. There are in places notable fans, subordinate to the apron as a whole. This apron by no means presents a continuous front, as it has been somewhat dissected since its formation. In some areas it resembles nothing so much as a series of parallel 'nine dumps; narrow, elongated, flat-topped, steep-sided, abruptly terminated. The sand and gravel between the "dumps" is of like origin ; the river alluvium appears at the base of the "dump." It is a strikingly notable feature that the outer edge of the apron preserves a constant level throughout its extent along the mountains. There are no intersecting slopes such as are so frequently found in areas having fan forms. Naturally, the drainage pattern of the area is dominated by the Colorado. There are represented in the area all the features of a stream having great seasonal fluctuation in the volume of water carried : natural levees, annual overflows, and alternate scouring and deposition in channels. There are the features of delta areas : delta lakes, anasto-
Fig. 2. Profile of the Colorado Delta from the Salton Sink to the Gulf of California.
154 University of California Publications in Geography [VoL. 5 mosing and distributary channels. Where river current is combined with the tidal action of the sea there are mud flats, ephemeral islands, and overflow channels which lead away from the main river channel. Within recent times the area of the river's activity has been largely restricted to the region lying between the delta crest and the gulf. Occasional overflows have carried water into Salton Sink and Pattie Basin, but these are mere incidents and the streampatterns developed in these areas are insignificant. The present deep channels of the New and Alamo rivers were culturally induced. No streams originate within the area. In fact, the mesa plains and large areas of the alluvial plains betray no signs of stream channels. Leading out of the canyons of Sierra Juarez are dry stream courses which, under the most favorable circumstances, carry water to the edge of the alluvium; but the necessary conditions occur infrequently. To a lesser extent this also holds for the desert ranges. Within the Cucopas there is a well-developed drainage pattern : waterfalls, pools, deep abrasion forms--all normally dry. The natural vegetation of the area shows striking contrasts both as to type and as to luxuriance. The abundant water of the Colorado has served to effect a marked dissimilarity in flora as between delta and surrounding desert. In areas of regular overflow or near permanent channels the line of distinction between delta and desert flora approximates the boundary between river silt and mesa sands and gravels. In other areas, the factors of alkali deposits and irregular water supply
rob this line of its significance. That portion of the delta lying south of the crest.is covered by a heavy growth of deciduous trees: cottonwood, black willow, mesquite, and sycamore. In the same general association appear the tules, rushes, and arrowweed. The old permanent stream channels are marked by the cottonwoods which line their banks. The lower lying overflow land is covered with willow, the higher with arrowweed. The latter grows in great profusion and frequently forms nearly impenetrable thickets. Mesquite is widely distributed and forms solid woodlands on the higher ground. About the sites of springs along the base of Santa Clara Mesa, in swampy sections of the lower delta, tules and rushes are found. These frequently occur in areas where other vegetation is lacking, so that their bright green color is refreshing in an otherwise drab landscape. On the islands of the river mouth, on the alluvial plains of the lower river, and about the brackish springs which border the moun-
1932] Kniffen: The Natural Landscape of the Colorado Delta 155 tains are patches of salt grass, frequently forming closed stands. About the margins of undrained depressions, the alkali lands, halophytic saltbushes stand as the vegetational outliers. The bed of Laguna Salada and the great alluvial plain stretching south from Sierra Mayor are without any cover or vegetation. Most characteristic of the mesa is the creosote bush. Accompanying it is the ocotillo with garambullo (Opuntia sp.) and cholla, appearing on the little sand hills. In the sandy washes appear palo verde, desert willow, and the smoke tree. The evergreen, parasitic. mistletoe appears in drooping clusters on the palo verde, palo fierro, and mesquite. On the detrital slopes is found the palo fierro, with visnaga growing on seemingly bare rock surfaces. On the higher portions of the Sierra Cucopas, and against the base of the Juarez, the maguey is found--though its restriction to these areas may be partly the result of exploitation by mescaleros. Deep in the canyons which lead from the Juarez, in at least one high valley of the Cucopas, a few Washingtonia palms grow, indigenous to the area and marking the sites of permanent water. These are the material elements of which the milieu is composed, but the picture is bare and lifeless without an appreciation of those other elements impossible to grasp and dissect at one's leisure. To make the scene vivid it is necessary to feel the scorching heat of midsummer, to follow the ever retreating mirage, to see the sky smoky with the great dust whirls of the alluvial plain, to hear the cry of the water fowl as the tide enters the river mouth, to see the sunset over the Sierra Juarez with the desert ranges as sharply distinct as a cameo. The cultural scene is restricted to the area lying north of Cerro Prieto and east of the Cucopas. Here the irrigation canals and the cotton fields of Baja California are a part of, yet distinct from, the agricultural development to the north of the border in Imperial Valley. South of this region the natural landscape remains effectively untouched. Several unfenced cattle ranches occupy the area; the cattle harvest the native vegetation. The flood, which brings to the agricultural district the serious problem of levee building, provides the feed for the stock of the cattlemen.
156 University of California Publications in Geography [Vol.. 5 THE TECTONIC FRAME The Peninsula of Lower California and the Gulf of California are generally explained in terms of the depression of an elongated section of the earth's crust lying between parallel fault lines, or at least bounded unilaterally by a fault zone
. For these forms the term graben is to be preferred to the term rift which is also sometimes used. The choice is only between a faulted surface and a geosynclinal one. Lawson, in writing of the area under discussion, said in 1908: The Colorado Desert and its continuation in the Gulf of California are certainly diastrophic depressions, and may with much plausibility be regarded as a great Rift valley of even greater magnitude than the now famous prototype first recognized by Suess. This great depression lies between the Peninsula of Lower California and the Mexican Plateau. All three of these features find their counterpart in southern Mexico. The Sierra Madre del Sur is the analogue of the peninsular ridge; it lies on the line of its prolongation, and is similarly constituted geologically.3 Freudenberg is in agreement with Lawson's views. He considers the whole structure to be a tectonic feature of the first order, and compares it with the Sacramento-San Joaquin Valley of California.' Whatever the gross structure of the feature may be cannot as yet be decided; in detail the whole area in recent geologic periods has been characterized by great changes in level. The most notable of recent episodes in its structural history seems to have been a very general and considerable uplift. The evidence is so plain and convincing as to have attracted the attention of nearly all, if not of all, trained observers who have seen the area. That the Salton Basin is a continuation of the structural trough of the Gulf of California is so evident as hardly to need mentioning. Suffice it to say that the deepest well drilled into the delta, to a depth of 400 feet below sea level, revealed only river sands and gravels, while at Coachella a drill hole showed no bedrock 1000 feet below the sea. 3 A. C. Lawson, The California earthquake of April 18, 1906. Report of the state earthquake investigation commission (Washington, Carnegie Inst., 1908), 1:52. 4 W. Freudenberg, Geologic von Mexiko (Berlin, 1921) :8-9.
] 1932 Kniffen: The Natural Landscape of the Colorado Delta 157 Though the Gulf-Salton trough does not continue as such to the north there is a northward continuation of a number of fault lines which may in part be responsible for its structure.° The San Andreas rift has been traced without a break from the Coachella Valley to the San Francisco
Peninsula. The San Jacinto fault has been tentatively extended by Beal to the mud volcanoes near Cerro Prieto, south of the international boundary.° Both according to local tradition and in some measure to official report, every major earth movement occurring in southern California has had its counterpart in local quakes and in increased activity of the mud volcanoes. Viewing broadly the general structure of southern California and the northern section of Baja California, there appear to be two directions along which the major structures trend. One trends somewhat west of north, the other northwest and southeast. The first marks the strike of the Sierra Nevada
of California, terminated by the Tehachapis on the south, and the Juarez-San Pedro Martir system in Baja California. Between these two ranges is an area in which the north-south lines are intersected by those striking northwest-southeast. Together with a third set of east-west lines these account for the striking and characteristic development of the series of structures previously described as spurs and reentrants, and applied to such features as the Santa Rosas and Borrego Valley. As suggested above, the Juarez-San Pedro Martir Range is in its structure--and, it may be added, in its lithologie character--very similar to the Sierra Nevada and somewhat contrasted to the mountain systems lying between.7 This range rises to much greater heights than do the ranges immediately to the north, it possesses a continuous and regular east-facing scarp, and its summit is plateau-like. Within the delta the prevalence of tectonic activity is emphasized by the frequency and number of earth tremors. There occur shocks of major intensity, but more frequent are minor ones felt in closely restricted areas. Of the former class Beal mentions those of 1852, 1892, 1915,8 and that of 1927 should also be included. Of the latter 5 The position of the major fault lines of California is shown on the maps accompanying the report of the California Earthquake Investigation Commission. Carl H. Beal, The earthquake in the Imperial Valley, California, June 22, 1915, Bull. Seis. Soc. Am., 5:130-149 (1915). 7The similarity between the Peninsular Range in Baja California and the Sierra Nevada of California is brought out in: S. F. Emmons and G. P. Merrill, Geologic sketch of Lower California, Bull. Geol. Soc., Am., 5:489-514 (1894), Op. cit., 138.
158 University of California Publications in Geography [Vor.. 5 class there are so many as to receive scant attention from the local inhabitants. A noteworthy feature of these minor shocks is that they may be felt at ohe point while a few miles away they are not detected. If the tremors represent displacements in the bedrock, lying far below the surface, it is to be expected that the alluvial cover should hardly reveal a distinguishable rupture. Such is generally the case, but on occasion there are surface breaks. One such, occurring perhaps forty years ago, was of sufficient magnitude to cause the diversion of the Colorado at a point near its junction with the Hardy. Another was traceable from Cerro Prieto along a line leading southeast through the Santa Clara Slough, and its course was marked by mud eruptions. Though no accurate record has been maintained, there is an impression prevalent among the engineers of the area that the frequency of the minor tremors is greatest at the time of the annual overflows, that is, in the summer months. At this time of year the added weight of the flood waters with their burden of silt is undoubtedly great, but this could hardly account for deep-seated movements. Possibly it is effective in causing superficial readjustments in the alluvium. Representative of volcanic activity in the area are a number of surface forms ranging in size from the extinct volcano of Cerro Prieto to active mud volcanoes, fumaroles, and hot springs. One group of mud volcanoes is found seven miles west of Niland, and in the same neighborhood are several obsidian buttes. The other group of active forms is found near Cerro Prieto (pl. 19a). Relict forms are to be found north of Cerro Prieto, along the Santa Clara Slough, and near Pozo Cenizo at the northern tip of Sierra Tinajas (pl. 19b). Isolated hot springs, in part sulphuretted, are found along the edge of Santa Clara Mesa, south of Mesa Andrade. The distribution of these forms is by no means haphazard, and taken in conjunction with other types of forms they would seem to lie along well-defined lines of weakness. The association of increased activity of the mud volcanoes with major earthquakes is now well established. This was noted by Major Heintzelman at the time of the earthquake of 1852 when clouds of steam rising from the volcanoes of Cerro Prieto were clearly observable at Fort Yuma, forty miles away.° 9 Reported to Professor Blake and cited by him in: Reports of explorations and surveys to ascertain the most practicable route for a railroad from the Mississippi River to the Pacific Ocean, made in 1853-4 (Washington, 1856), 5:115.
] 1932 Kniff en: The Natural Landscape of the Colorado Delta 159 When particularly active they make a roaring and booming noise which is said to resemble the sound of heavy guns. The major relief features of the area owe their position as such to the agencies of diastrophism and volcanism. These endogenous forces have created the framework about which the secondary, exogenous forces produce the complex of forms which constitute the natural landscape. As the two sets of forces work simultaneously there are few forms which are entirely the work of endogenous forces, unmodified by the forces of weathering. Only some of the minor volcanic forms are exceptions. The fact, however, that relief exists proves that the reliefcreating forces have at some time acted at a rate faster than that of the relief-destroying ones. The slope relations within the attacked form and between it and the depositional form reveal the history of the relative dominance of one or the other set of forces. The whole interrelationship makes a descriptive classification of "tectonic" or "volcanic" forms quite difficult, so it is here chosen to make a classification of "relief forms, particularly forms of positive relief having their origin in diastrophism or volcanism. This classification is descriptive, is based largely on the magnitude of the forms, and so disregards the sequence of events which gave them their relief. DIASTROPHIC AND VOLCANIC RELIEF FORMS First order: In this class, standing alone, is the great depression which is the trough of the Gulf and Salton Basin. This also includes the depression of Pattie Basin (Laguna Salada). Second order: Sierra Juarez--San Pedro Martir, particularly the east-facing scarp. Granitic face varying in slope from thirty to ninety degrees from the horizontal. Sufficiently steep to prevent the accumulation of residual or detrital material, therefore largely vegetationless. Possibly to be put in the preceding class as a constituent of the Gulf trough. Third order: Sierra Cucopas. Dominantly a granitic mass, yet containing metamorphics and volcanies. Eastern face somewhat steeper than the western. Slopes varying from perhaps twenty to ninety degrees. In places sufficiently gentle to permit the accumulation of detritus. Sierra Las Pintas. Very irregular in form. Volcanic in origin, made up of tuffs and lavas, in part bedded. Slopes from zero to ninety degrees. Sierra Las Tinajas. More regular in form than S. Pintas. Crystalline rocks present but strongly dominated by bedded lavas and tuffs. The northern portion a basalt capped mesa. Slopes of all degrees. Fourth order: Volcanoes such as Cerro Prieto and the isolated forms bordering the Sierra de las Pintas. Essentially flows, yet containing some cinder and
160 University of Calif ornia Publications in Geography [Vol.. 5 ash. Slopes not exceeding forty-five degrees, and many sufficiently low to permit of accumulations of detrital material. Fifth order: The minor volcanic forms such as the mud volcanoes, both active and extinct. The mud contains siliceous elements, the relict forms are resistant to the forces of weathering, and tend to retain the steep slopes of the active forms. In addition to the above classified forms, there are others which, as relief features, may be tectonic in origin. Notable among these are certain areas of high ground in the lower delta which have long been free from inundation by the annual floods. Then there are forms which owe their being indirectly to tectonic activity. Mesa Andrade is quite certainly a former portion of Mesa Santa Clara, which has been cut off by erosion along a fault line. The course of New River follows a fault line in part. The river terraces and the ancient beach line represent indirect tectonic forms. With this consideration of the tectonic and associated volcanic forms, by fixing them on a map and noting their relative positions, it is possible to detect what appear to be certain well-established lines of weakness. If topography is a criterion, one of these lines passes along the eastern front of the Sierra Juarez. Another is certainly indicated along the eastern front of Sierra Mayor (pl. 19c). That this extends continuously along the whole front of the Mayor-Cucopa system is by no means so certain. A third, better established line is found in the southward extension of the San Jacinto fault. As previously cited, Beal tentatively extended this to the mud volcanoes near Cerro Prieto. It is proposed to extend this in the same general direction passing it to the east of Mesa Andrade, along the course of the Santa Clara Slough, and along the base of Santa Clara Mesa. South from here its course is rather uncertain. It may continue to the southward along the coast or it may pass inland and account for the sudden appearance of the Tertiary sandstone mass which rises above the Quaternary gravel plain of Mesa Santa Clara. A careful working out of the complex geologic history of this area is beyond the scope and intent of this report. A brief resume of the principal geologic events, particularly with regard to their magnitude and sequence, will provide, however, the necessary orientation before proceeding to a consideration of the exogenous forces. The following is a simple geologic classification of the rocks and unconsolidated materials which form the principal surface exposures in the area. An understanding of their geologic time relations and an
1932 ] Kniff en : The Natural Landscape of the Colorado Delta 161 examination of the topographic positions which they occupy with respect to each other serve as a key to the working out of the earlier tectonic history of the region : 1. Pre-Tertiary granites. The Peninsular Range and the granitic desert ranges are composed principally of igneous rocks belonging to this group. 2. Tertiary sediments. Both consolidated and unconsolidated sands, clays, and gravels. Both marine and terrestrial. Generally highly disturbed and in places slightly metamorphosed. They underlie the Quaternary and Recent deposits which form the surface of the Salton Sink. They are found in surface exposures along the base of the Peninsular Mountains, in places along the Cucopas, near Cerro Prieto, and east of the head of the gulf. Near the Salton Sea they form a part of the "mesas" and exhibit beveling. 3. Late Tertiary extrusives. Basalts, tuffs, cinder. Bedded flows and volcanoes. Cerro Prieto and Sierra de las Pintas. 4. Quaternary sands and gravels. The principal constituent material of the mesas. Largely undisturbed and exhibiting marine planation. 5. Recent alluvium. Largely the material derived from the Colorado and forming the surface of the delta. By middle Tertiary time the gulf trough seems to have been fairly well established as a structural feature. The sea stood hundreds of feet higher than it does at present and extended far beyond the present head of the gulf. The Peninsular Mountains and probably the Cucopas were islands. From the mountains was derived the material for heavy marine deposits. Conditions were favorable to the existence of abundant marine life forms. With late Tertiary there was a decided uplift. Quite possibly this was accompanied by a depression of the gulf trough. At least there was enough differential movement so that the earlier Tertiary sediments along the
162 University of California Publications in Geography [ VOL. 5 base of the Peninsular Mountains were highly distorted ; some of them were faulted. During this same period there was very considerable igneous activity, and it probably continued into early Quaternary. This activity resulted in the formation of the Sierra Pintas and such isolated forms as Cerro Prieto. By middle Quaternary the sea had advanced again and stood at an elevation of perhaps a hundred feet above present sea level. It reached to the base of the mountains and it lay over the mesas ; to some extent it beveled off the folded and faulted Tertiary deposits. During the Recent period there has been a general and fairly evenly distributed uplift--as demonstrated by the old beach line cut into Quaternary and Recent sediments, and by the fact that the Quaternary sediments are little disturbed. The most significant depositional activity has been the continuation of valley filling initiated in the Tertiary, principally through the agency of the Colorado River and its alluvial burden (fig. 3). These are the raw materials with which the exogenous forces have had to deal. The diastrophic and volcanic forces have created a framework ; the river has contributed its alluvium. Out of all this the unique ensemble of exogenous forces, characteristic of this particular area, has created a set of surface forms. RIVER AND SEA THE RIVER The watershed of the Colorado system includes parts of five western states and comprises a total drainage area of some 265,000 square miles. Most pertinent to a study of the delta region is the fact that climatic conditions within the area of the watershed vary between wide limits. The course of the river below the Grand Canyon
passes through a region of increasing aridity with winter temperatures above freezing. The area of greatest precipitation lies about the headwaters of the Grand in Western Colorado. Even here the precipitation does not exceed thirty inches but the effective run-off is great, as this is an area of cold winters and of heavy winter snows ; even in summer the initial loss due to evaporation is not great. Here, as on the headwaters of the Green, small glaciers, snowbanks which do not melt until late summer, little lakes, and springs occur. It is these climatic conditions, modified by the nature of the relief, soil cover, and vegetation cover that determine the regimen of the
1932] Kniffen: The Natural Landscape of the Colorado Delta 163 Colorado near its mouth. Of those tributaries above and including the San Juan it may be said that they are dependable and that the volume and seasonal fluctuation of the water discharged by them are predictable. In the spring the flood comes as the result of the melting snow ; throughout the year these streams furnish the bulk of the water. The lower tributaries are unreliable; during much of the year they contribute nothing; at unexpected times they contribute sudden "flash" floods of large volume whose effectiveness as transporting and erosive agents is very considerable. Of these latter streams may be mentioned the Gila, Bill Williams, Virgin, and Little Colorado. In the anomalous relation which exists between its lower course and the climatic region in which it lies, the Colorado presents a phenomenon similar to that of the Nile. The latter stream also rises in a humid area and has its mouth in an area in which the stream's loss in volume is greater than its increment. That such a relationship should produce unique results in the landscape has already been suggested and will be further developed in a later chapter. An examination of the twenty-four year average discharge of the Colorado as recorded at the Yuma gauging station reveals a number of points. From September 30 to March 30 the discharge remains fairly constant with an average volume around 13,000 second feet. The average maximum occurs during the last week in June and reaches approximately 81,000 second feet." Of course, this represents "average" conditions. It does not show the occasional "flash" floods such as that of February, 1920, when the Gila brought the Yuma discharge up to 164,000 second feet for a day. Nor does it show the fact that very frequently the annual summer flood occurs not in one peak but perhaps in two, three, or four maxima, which may come as they did in 1926, at monthly intervals, the first on April 12, the fourth on July 19. Still it remains true that the river floods its banks in the summer months and is low during the winter months. The amount"- of absolute maximum and minimum discharge are : absolute maximum, 186,000 second feet, June 28, 1921; absolute minimum, 1,200 second feet, September 11, 1924 (fig. 4). There is at Yuma a mean annual run-off of 17,000,000 acre-feet." Translated into comparative terms this means one-fourth of the run-off of the Nile at Cairo, and in general this relationship is maintained between the Colorado and Nile as to run-off, length, and irrigable area. 10 These and other figures were furnished by the Yuma office of the U. S. Reclamation Service. "E. C. La Rue, Colorado River and its utilization, U. S. Geol. Sum, Water Supply Paper 395:23 (1916).
164 University of California Publications in Geography [Vol. 5
The lower valley of the Colorado above its delta is irregular in
form, this irregularity being dependent upon the proximity of the
mountain ranges which bound it on both sides. As a result, the width
10 0 .t I 00 115 110 '44:) sco et· IT, goo 44Z. 05
- -3A eaAR-
Bauraam C7t1JJ (Section Colorado Riwr- JuneP. 8, 1921
· Sept 11, 1924
Fig. 4. Cross-sections of the Colorado River at Yuma at extreme low water and at flood.
of the valley varies from the width of the river in gorge stretches to perhaps thirty miles, producing in effect a series of subsidiary valleys connected by the course of the stream. As suggested, the minor tributaries are few and insignificant. 12 This whole matter is treated so well in other sources as to require only the BbFrruairenyfc.e,iWssctoar,tee1fre9r1Se5un)pc;peElyh. ePCra.eLp. aeSreReu3H9e5,. CT(o1. 9lCo1or6ra)yd, oIamnRdpiveWeriraatalenVrdaploiltewsyeuartnialdnizdaStafilolotnoo,ndUcS.oiSnn.ktGr(oeSloaonlf. Colorado River below Green River, Utah, U. S. Geol. SURV., Water Supply Paper 556 (1925).
1932] Kniffen: The Natural Landscape of the Colorado Delta 165 But what is most significant about this valley for the purposes of this paper is the fact that the Colorado in its lower course is not a bedrock stream and that its flood plain is bounded on both sides by a continuous terrace. Soundings indicate that in all the canyons of the lower river bedrock is at least one hundred feet below the river bed." The terrace is marked by a conspicuous bluff which rises from fifty to one hundred feet above the flood plain and continues far to the north along the valley." From these conditions the natural inference is that the valley has been subjected to a deep fill and that subsequent to this event the cutting and transporting power of the stream has been markedly increased so as to cause the formation of terraces. Further reference to these events and their implications for the delta area will appear later. The silt burden-- Though the river as such is a feature and form of the landscape, its tremendous geographical importance lies in its present and past activity as an agent or tool in creating surface forms out of its own soil burden. This alluvial matter, though it varies in size from a very fine sand to colloidal clay, is commonly called silt and will here be so designated. As a carrier of silt the Colorado is probably without a peer among the greater streams of the world. An examination of the river water at Yuma over a period of several months in 1892-93 revealed an average ratio of 1 to 277 for dry material to water. Corresponding ratios for the Nile would be 1 to 1900; for the Mississippi, 1 to 1500; for the Danube, 1 to 3060." In 1904, Forbes" of the Arizona Experiment Station made a careful study of the river silt. He found that an acre-foot of Colorado River water contained on an average 9.62 tons of silt, and that for the year the river's burden amounted to over 120,000,000 tons--this for a year when the total discharge was considerably under normal. The average annual load passing Yuma is probably around 160,000,000 tons, which translated into terms of volume of dry soil would be approximately 80,000 acre-feet'? 13 W. T. Lee, Geologic reconnaissance of a part of Western Arizona, U. S. Geol. Sum, Bull. 352:66 (1908). 14 J. S. Brown, The Salton Sea region, California, U. S. Geol. Bury., Water Supply Paper 497:34 (1923). 13 C. B. COLLINGWOOD, Soils and waters, Univ. Ariz. AGR. Exp. Sta., Bull. 6:7 (1892). 16 R. H. Forbes, The river irrigating waters of Arizona--their character and effects, ibid., Bull. 44:200 (1902). 11 E. C. La Rue, U. S. Geol. Sum, Water Supply Paper 395:220.
166 University of California Publications in Geography [Vol.- 5 It is to be expected that the silt burden of the stream should vary at different times of the year, and this is true. An increasing volume of water, however, does not mean proportional increase in silt content. This is to be explained in part by the fact that a falling stream by bank cutting and caving is a more efficient remover of silt than a rising stream, in part by the fact that the lower, "erratic" tributaries of the Colorado contribute a proportionally greater amount of silt than the upper, "reliable" ones. In this connection it may be added that each tributary contributes silt so unique and characteristic that those familiar with the area claim the ability to recognize the source of the deposited delta alluvium by its color and consistency. The lower tributaries of the Colorado, namely, the Gila, Bill Williams Fork, Little Colorado, and San Juan carry more silt during their flowing season than does the upper Colorado." They are largely the effective causes of the "flash" floods of spring and autumn so that the silt burden in these periods is proportionally greater than that of the much larger summer flood." These lower tributaries contribute most of the true silt and colloidal material. From the Virgin to its headwaters, the Colorado, even in high water, contains a fine, siliceous material commonly called quicksand.2° Whatever the origin of the material, the essential point is that the Colorado below Yuma is a stream containing an enormous silt burden, that the silt is not only the passive material from which the delta is formed but that by its presence it introduces important modifications in the mechanics of stream flow. Mechanics of stream flow-- It is pertinent to mention here a few peculiarities of a silt-burdened stream, such as the Colorado. It has been said by an engineer working in the area that the silt tables generally accepted as applying to siltbearing streams fail of accuracy for the Colorado Delta area. Whether or not this be true the presence of silt in the stream produces some very interesting phenomena. 18 F. L. Sellew, in Cory, op. cit., 1480. 18 An excellent digest of the available data concerning the Colorado basin with particular reference to the relation of rainfall, run-off, and evaporation with a section devoted to areal distribution of denudation, is contained in E. Reichel, Der Wasserhaushalt des Coloradogebietes im siidwestlichen Nordamerika, mit einem Anhang, Untersuchungen iiber die Denudation dieses Gebietes, von Sr. M. Mercedes Leiter. Geographische Abhandlungen hrog. v. A. Penck, tier. 2, no. 4 (Stuttgart, 1928). 20 C. E. Grunsky, in Cory, op. cit., 1535.
1932] Kniffen: The Natural Landscape of the Colorado Delta 167 In its lower course the Colorado is alternately a degrading and an aggrading stream. This is strikingly brought out by sections taken of the channel at Yuma during periods of high and low water (see fig. 4). Obviously the stream becomes degrading after its volume has reached a critical point ; but this is not totally dependent on the increased carrying capacity of the larger amount of water. There may appear the paradox of an increasing discharge and a falling gauge height. An examination of the Yuma gauging records for a period of years showed that the cross-section in a given year increases from minimum to maximum ; 39 per cent of the increase is due to rise in the water level, 61 per cent to scour. The critical point above which the river scours and below which it deposits seems to be around 47,000 to 50,000 second feet--though this figure is by no means a fixed one. A part of this scouring goes on at a time when the river water is apparently carrying in suspension its greatest possible burden of silt. The continued scouring beyond this point has been explained as the bodily pushing down stream of the silt resting on the stream bottom by the hydrostatic pressure exerted by the superimposed mass of water and suspended silt. Within the bed of the stream where the course is straight the lowest point of the cross-section seems to seek first one side and then the other. This feature is reproduced in the irrigation ditches under ideal conditions of a straight channel and a constant head of water. Here the current scours on one side and deposits on the other, then conditions are reversed. In other words, these are incipient meanders which migrate slowly downstream. During high water, on the bends of the stream there is to be observed a phenomenon in which, with monotonous and periodic regularity, the main force of the current swings toward the ouside bank, suddenly advances in a wave which may reach two feet in height, then swings back toward the center of the channel to repeat the performance. This is probably to be accounted for by the building of bars from the inside of the bend. These bars increase in size until they act as a dam piling up a head on their upstream side. The head reaches a critical point, then advances, suddenly scouring out the bar. A natural bar, a weir, a sharp bend, the increased volume of a tributary introduce a measurable decrease in the velocity of the stream far above the obstruction. This then is followed by an increased velocity below the obstruction.
168 University of California Publications in Geography [ Vol-, 5 Of interest and certainly of importance to levee builders are the "bores" which may develop without warning at the bases of the levees. The water suddenly starts a whirl, and eats its way through the embankment, without any apparent seepage by which it may start. These are extremely difficult to stop and account for most of the levee breaks. They also occur along the bases of levees which have become dry and cracked. Before the artificial diversion of the Colorado from its old channel in 1905 the annual flooding of the delta area came from comparatively gentle, over-bank flow. There were occasional minor crevasses, but these were the exception rather than the rule. This type of overflow is still maintained on the lower delta. By this process, natural levees are built and added to from year to year so that the land slopes gently away from the stream. The coarser, sandy material is deposited immediately and marks the presence of old or contemporary channels. The deposited alluvium increases in fineness with distance away from the stream course. Where there are natural depressions, such as ox-bow lakes, which act as settling basins, there is deposited the finest material, forming an adobe. Those familiar with the region think they can identify the source of the deposited material: thus "red silt" is believed to be contributed by the Little Colorado, "black silt" by the Grand, etc. A part of the flood waters is lost by evaporation, certainly very little by seepage. The remainder spreads in a sheet for some distance from the main stream and then is gathered in by small channels which converge into larger ones. Eventually the waters are returned by a tributary to the main river. The secondary deltas-- The two preceding paragraphs give a schematic sketch of the development of the delta, particularly as it was before the diversion of the Colorado from its old channel. Since then the river, by processes not thoroughly understood, has constructed a new set of unusual forms. These are the secondary deltas with their anastomosing channels. In their formation the whole stream has been deflected from the old course and turned across the country where other channels are not available. On all this overflow land, brush grows in the greatest profusion, which serves as a check and causes the larger portion of the silt burden to be dropped. The water, without a channel and with
1932] Knifes: The Natural Landscape of the Colorado Delta 169 thick brush obstructing the way, is forced to find its way through, here and there, at the weakest points. On the other side of the brush it finds its way to one of the old gathering channels and becomes again a united stream. In January of 1922 the main stream, which was then flowing to the west through the Bee River channel, was diverted toward a branch of the Pescadero, one of the gathering channels of the Hardy. The brush was cleared and a channel dug for a distance of several miles. The slopes away from the channel were at the rate of sixteen feet in five miles. After the first summer flood this had increased to five feet to the mile. The dug channel had been scoured by the flood waters to a depth of twenty feet. The greater part of the silt had been deposited within five miles of the end of the constructed cut; beyond that there was a fall of six feet in two hundred feet distance, then an open channel, and beyond that were small braided channels with rapids and log jams. In all this course only adobe was deposited, no sand or sa,ndy loam. By the fall of 1926 the river had canalized itself for a distance of six miles beyond the end of the constructed channel and was continuing the process at the rate of two miles each year. This it was doing by constructing a delta cone and extending it. Each year silt was deposited to a depth of thirteen feet immediately at the end of the channeled portion of the stream. From this the cone sloped off laterally to no deposit at a distance of 4000 feet. The stream constructed its channel over the surface of the old delta and after it was established turned about and scoured, to some extent, its own self-constructed bed 21 These brief references to some of the more striking processes whereby the river constructs its forms will become more coherent in the description and classification of the forms of the delta which owe their origin to the river and the deposition of its silt burden. The forms-- By way of orientation for the sections which follow it may be well to insert here a brief description of the pattern which the stream channels have developed within the delta. Reference is made to the map where the details are expressed graphically. 21 The facts on which these two paragraphs are based were taken from two unpublished manuscripts by R. M. Priest, U. S. Reclamation Service engineer at Yuma, and kindly made available by him.
170 University of California Publications in Geography [VoL. 5 With the exception of the period 190547, the main channel of the Colorado lay along the eastern margin of the delta until the year 1909. Apparently equally stable as a feature of the delta was the channel of the Hardy which occupies a position along the western margin of the delta, and at that time extended northward to Volcano Lake. South of the crest the general slope of the delta was away from the Colorado and toward the Hardy. The main stream was confined to the old channel by natural levees. During the overflow period water passed over the banks and as it moved slowly westward was gathered into small channels such the Bee, Pescadero, Paredones, and Alamo. The first three of these were tributary either to Volcano Lake or the Hardy, and most of the water which they carried eventually found its way to the gulf through the Hardy. The Alamo lay north of the crest and carried its waters to the Salton Sink. A portion of the water entering Volcano Lake found its way through New River to the Salton Sink. From 1905 to 1907 the main stream of the Colorado flowed through the New and Alamo to the Salton Sea. In 1907 it was returned to its old channel, and flowed to the gulf until 1909. From 1909 until 1922 the main channel was through the Bee River to Volcano Lake, from which most of the water reached the gulf through the Hardy. By means of a dam, the main channel was diverted from the Bee to a tributary of the Pescadero in 1922. Since that year the Pescadero has remained the main stream, the water passing into the lower Hardy and so to the gulf. Properly to be considered first are the normal river channels with their associated features, eliminating for the present those very special types which are particularly ephemeral. The channels immediately fall into two general classes : those which are cut through the clays or adobes and true silts, with resulting narrow stream courses and steep, not easily eroded banks ; and those which are cut through the coarser, more easily eroded soils where the banks are less steep and the streams meander and cut new channels more easily '2 This differentiation, which obviously has its basis in difference of soil type, becomes largely one of proximity to the old bed of the Colorado. Such streams as the Hardy, Pescadero, New, Alamo, and the mouth of the Colorado fall into the first group, those having steep banks. The upper part of the old channel of the Colorado, and most of the Bee and Paredones, fall into the area of coarser, sandy soils. 22 This has been previously noted by Godfrey Sykes, an excellent observer, long familiar with the area. Geog. Rev., 16:250 (1926).
1932] Kniffen: The Natural Landscape of the Colorado Delta 171 What might be termed "channel" forms are the islands, sloughs, bayous, and lakes whose origin has been due to meandering (pl. 20a). These forms are present in greater or lesser number over all the area, but here again there is a difference, according to the soil through which the channel takes its course. In the areas of sandy soils they are more abundant but also more ephemeral. In the areas of adobe soils they are fewer in number but tend to be comparatively permanent. Certain lagoons and islands of the lower delta have been used by the Indians since time immemorial and are still in existence. Bounding the upper part of the delta, and extending up both the Colorado and Gila, are the terraces previously alluded to. Though no material coarser than a very fine sand appears in the delta deposits, these terraces are composed of coarse sand and gravel. They rise abruptly from the flood plain of the river and extend as gravel and sand plains to the mountains. They vary from thirty to fifty feet in height above the flood plain, and their absolute elevation varies with the gradient of the streams which they bound.28 A section through the terraces reveals alternate layers of sand and gravel with the usual features of cross-bedding and lenticular inclusions (pl. 20b). Of great importance in their implications for surface forms are the underground channels and general sub-surface water conditions. The major source of the underground water
is, of course, the Colorado; only about the desert margins or in extremely deep-seated horizons are other sources significant. Within the delta the conditions which determine the presence or absence of ground water are those of soil texture. The finer colloidal material forms a very effective barrier to the percolation of water. After rains or overflows the water stands in pools until removed by evaporation. In the western part of the delta, composed as it is of silt and colloidal material, the water table is low, except along the banks of permanent channels. In such areas as the Yuma Valley, where the underlying soil is a quicksand, the water table is high, having approximately the same elevation as the river. Where the delta borders on the mountains, as along the base of the Cucopas, mountain-derived sands underlie the river alluvium and are water bearing. The above observations concern themselves with such surficial conditions as are significant in determining the presence or absence and abundance of plant growth. With reference to deep-seated waters 22C. P. Ross, The Lower Gila Region, Arizona, U. S. Geol. Surv., Water Supply Paper 498:75 (1923).
University of California Publications in Geography
there is a variety of conditions. Within the Salton Basin two areas of deep wells, in part artesian, are developed. In the Coachella Valley the mountain-derived sands and gravels form the water-bearing strata, the overlying silt and clay belts the impervious stratum, which together produce artesian conditions. Figtree John Spring and Fish Springs are natural indicators of this condition. In this area the ground water has its origin in the mountainous region which bounds it.24 In the eastern portion of Imperial Valley, about Holtville, there is another artesian area whose waters are derived from the Colorado. This area seems to he strictly limited in extent. Here the wells have been drilled to a distance averaging several hundred feet to a stratum of sand or gravel. A few miles to the west, holes drilled much deeper are dry or yield very little water." South of the boundary an area of pumped wells along the delta crest suggests an underground condition as yet not thoroughly understood. Well logs reveal a gravel "ridge," lying about two hundred feet below the surface, which extends in a direction along the delta crest, west by south from Pilot Knob. In shape this appears to conform closely to the delta surface, thinning out on both sides of the crest line. The gravel is fairly well rounded igneous material, said to be derived from the Gila. There seems little possibility that the origin of the water could be any other than the Colorado, though its fluctuation of level follows the fluctuations of the Colorado only in minor degree. It seems possible that this water, as well as that of the Holtville wells, might very well be derived from the continual loss of the Colorado between Grand Canyon and Yuma, suffered gven during the low water stage of the river.
Overflow forms-- Another set of forms is introduced by the overflows in which the river leaves its established channel and passes out over the surface of the delta. First in point of origin are the natural levees. Generally these are distinguishable only by instrumental observation as their slope away from the river is ordinarily gentle, say at the rate of five feet to the mile. At certain stages of the overflow period the levees stand out by virtue of their covering of vegetation. The overflow plains are a part of the levee slope. On these the flood waters lie in a sheet, not yet gathered into the channels which 24 J. S. Brown, U.S. Geol. Surv., Water Supply Paper 497:73-75. 25 /bid., 78-83.
1 ]932 Kniffen: The Natural Landscape of the Colorado Delta 173 restore them to the main stream. These forms are characterized by a gentle slope away from the stream channel, are dissected only by minor gathering channels, and vary in material from the fine sand of the main channel to the silt of their outer margins. With these there are also associated small lagoons which represent portions of old channels or ox-bow lakes. The lagoons may be without inlet or outlet channels, and may in time become exterminated through silting. Volcano Lake, Laguna Magnate, and the Salton Sea represent cases which will be considered later. Dissecting the outer margin of the overflow plain are the gathering channels--narrow, steep-walled channels as compared with the main stream. The description of the above forms is based upon conditions as they were before the Colorado was diverted from its old channel. The present condition of secondary delta building, with its accompanying phenomenon of anastomosing channels, probably represents the normal conditions by which that portion of the delta lying south of the crest has been built. These forms, then, have had their counterparts in the past history of the delta. Under conditions whereby the main channel builds itself up to a height above that of the surrounding country, it is inevitable that eventually it should swing out to a more favorable gradient, establish a new channel by gradually extending a secondary delta, and then repeat the whole performance. With the retreat of the flood waters several minor features are formed in the deposited silt. Probably the most striking of these are the mud cracks. These reach their greatest development where there has been a thick deposit of the finest alluvium. The surface breaks into large, irregular sections which are separated by cracks that may reach a width of several inches and extend downward to a depth of several feet (pl. 20c). If the deposit is thinner or composed in part of coarser material the sections are smaller and more superficial, the edges curl up and they readily disappear through crumbling (pl. 21a). Associated with the larger mud cracks are small cone-shaped depressions of perhaps twelve inches in diameter by twelve to fourteen inches in depth which apparently represent a point where the covering waters have found an opportunity to escape to a more permeable stratum below (pl. 21b). These forms were seen only on high water portions of the channel itself. In addition there should be mentioned the ripple and current marks which appear in the coarser material within a drying channel.
174 University of California Publications in Geography [VOL. 5 Influence on vegetation-- The river, indirectly, as the agent of soil differentiation, through its presence as water and as a bearer of seeds, creates, within the delta area, a characteristic vegetation pattern. Cottonwood, willow, arrowweed, mesquite, grass, and the various reeds are the main constituents of the vegetation formation whose areal variation in make-up is an expression of conditions imposed by the river. The cottonwood or alamo is largely restricted to a narrow margin along both banks of the old channel of the Colorado. Here it has a continual supply of water without suffering long periods of too abundant water, as is true of the flood plains on either side. This position also suggests the area of the coarser alluvium. The optimum conditions for mesquite demand water within reach of the roots without a considerable period when water covers the surface. As a result, the best stands are found on the finer soils occupying areas only occasionally flooded, or about the lagoons, or along such streams as the Hardy. Arrowweed occupies areas subject to overflow, but cannot stand long continued submergence. From its distribution it appears certain that its seeds are water-borne. It forms pure stands of great density in the overflow areas (pl. 21c). Willow demands and can stand a great deal of water. Its seeds are water-borne and it occurs in great thickets in such places as the necks of stream bends, where the soil is coarse in texture and is last exposed by the retreating flood waters. On the sloping banks of channels, willow sometimes is found in rows running parallel t; the course of the stream. The larger trees are found in the upper row and the other rows grade down to the size of the lower row. Each row evidently signifies a seeding at a certain stage in the river's height (pl. 22a). The rushes and tules occupy sites which provide permanent surface water. These conditions are found in the quieter portions of the channels, about lagoons, and in certain swampy areas of the lower river. Very great changes in the distribution of vegetation have been introduced by the abondoning of its old channel by the Colorado. According to one man, familiar with the area for fifty years, the country as he first knew it was one of extensive grassy plains. Cottonwood and willow were restricted to the main channel while arrow-
1932] Kniffen: The Natural Landscape of the Colorado Delta 175 weed and tule were found only in the area of the extreme upper delta. Mesquite was rather widely distributed over the delta, along the streams, and about the lagoons. This description corresponds quite well with recorded accounts of the last century.26 It is certain that there were large areas, such as that lying between the New River and Centinela, which were largely vegetationless. The Holmes map of 1902 contains some notations as to vegetation which bear this out. This is also largely true of the Silsbee" map of 1904, though this latter shows heavy willow timber east of Volcano Lake along the course of the Paredones. This may be significant, since by 1904 the Paredones had become in part an actual diversion of the Colorado to the westward. Old photographs show "savannas" with heavy grass stands, stands of guelite, canaigre, and carissou which are mentioned as good cattle feed, all conditions extremely hard to find at present. If this represents a true picture, certainly conditions have been greatly altered since the diversion of the river to the west. There has been a great extension of arrowweed and willow, and the grassy areas have largely disappeared. All this represents a change in natural conditions
and does not take into consideration changes effected by man which have been great and will be lasting. About the channels of the Bee and Paredones there are great thickets of arrowweed, and willow is extremely widespread. The lower delta remains least altered. Near the mouth there are areas of salt grass, growing under conditions which favor its survival. An exotic note is introduced by the coming of Bermuda grass, which has gained a firm hold on the lower delta within the last few years. Willow is present, as is also arrowweed, but they fail to dominate as they do farther upstream. The mesquite here holds its own. This whole matter of the alteration of vegetation seems to be, in this area, a question of the greater or lesser availability of water, and of changes in soil conditions. The river has changed its course, with marked results for the vegetation. 26 W. H. Emory, Report on the United States and Mexican Boundary Survey, 34 Cong., 1 seas., House Ex. Doc. 135, 1:105 (1857) (Mischler's report) 2:80 (Parry's report). 27 The originals of these two maps are in the possession of the Colorado River Land Co., for whom they were made. 28 quelite= Antaranthus pcanteri. canaigre= Rumex hymenosepalus. carisso= Phragmites phragmites. (Cane or tule.)
176 University of California Publications in Geography [Vol.. 5 TIIE SEA From a width of seventy miles at San Felipe, the Gulf of California narrows to the river mouth, only fifty miles north. An examination of the pilot chart shows a sharply bounded trough in the sea floor near the center of the gulf, east of San Felipe.29 This disappears farther north and near the head of the gulf the floor is comparatively level at a depth of about nine fathoms. Curious features are the occasional patches of red-colored water which appear in the upper gulf, which have been ascribed to the presence of river alluvium in the sea water. The true explanation lies in the presence of great numbers of flagellate infusoria, which float on the water and give it a milky-red color." Being without significant currents the temperature of the water follows closely that of the land about it, which means that it is fairly warm throughout the year. During the winter and spring months the prevailing winds are from the northwest, during the summer months from the southeast. This finds a simple explanation in the fact that the gulf, being a water body, preserves a fairly constant temperature through the year. The land surface about the head of the gulf varies considerably in temperature from winter to summer. This means that with respect to the land, the gulf becomes a low pressure area during the, winter and a high pressure area in the summer. In the upper gulf there are frequent northwest gales, lasting two or three days, during the months of December, January, and February. Winds of extreme severity are few though waterspouts are said to be of fairly frequent occurrence about the mouth of the Colorado. There is a decided land-and-sea breeze about the upper gulf, during the period of warming days in spring. Quite probably this phenomenon extends in lesser measure through other portions of the year. Hardy,91 sailing into the river mouth area in August, mentions storm waves of chaotic wildness, raised by the cross-seas, owing to suddenly shifting winds. 29 U. S. Hydrographic Office, Pilot Chart 619. so U. S. Hydrogr. Office, Mexico and Central America Pilot (west coast), ed. 7 ( Washington, 1928) :120. 31 R. W. H. Hardy, Travels in the Interior of Mexico, (London, 1829) :319.
1932] Kniff en: The Natural Landscape of the Colorado Delta 177 The upper gulf is sufficiently narrow so that under favorable wind conditions the surface of the water may become as calm as that of large lakes ; there is no constant surf. During the storm periods the waves strike the shore with sufficient height and force to be efficient tools for the cutting and carrying of the materials of which the strand is composed. Tides-- The whole gulf is so proportioned as to be favorable to the develop- ment of tides, and with the narrowing of the mouth of the Colorado conditions reach a climax which produces tides of unusual height and rapidity of change. At San Felipe, spring tides rise about twenty feet ; at Phillip 's Point in the mouth of the Colorado, there is a spring range of 31.5 feet, or an average range of 21.6 feet." The tides here sometimes run at a speed as high as six knots, and there is no slack water between ebb and flow. The most striking feature of the incoming tide at the river's mouth is the bore, or, as the Mexicans call it, "el burro." Without the warning of slack water, the ebbing tide is met by a wave extending from bank to bank and sweeping upstream. At spring tides this wave is normally about four feet high, and may rise to ten feet in shallow water. The distance upstream to which the bore, and the tides in general, are effective is dependent upon the stage of the water. When the river is very low there is a bore of one and one-half feet at Mayor, one hundred miles by river from Phillip 's Point in the mouth. When Volcano Lake contained water the tide was felt there, and at present the tidal effect extends up the Hardy to a point near the dry bed of the lake. In the old channel of the Colorado the tides could ordinarily be felt up to Heintzelman's Point. Again, the freshness or salinity of the water of the lower delta is dependent upon the river's stage. During flood periods the river water is potable even at half-tide as far down as Coyote Slough, opposite Montague Island. During periods of low water in the river the Hardy is scarcely potable up to Mayor. At one period of continual low water in the lower river small sharks and other salt water fish ascended nearly to Mayor and the water could not be drunk, even by cattle. 82 These figures are from p. 180 of the Mexico and Central America Pilot (west coast), previously cited.
178 University of California. Publications in Geography [Vol,. 5 Significant as the tides are as forces which have created characteristic forms in the confined channels lying within the sphere of their influence, they are of even greater importance as the force which meets the flood waters of the river and produces the overflows of the lower delta. The overflows of the upper delta come as a result of normal over-bank flow. In the lower delta the swollen river water meets the resistance of the incoming tides with a resultant building up of a head which finds relief in overflow. It is this flooding that, under favorable conditions, sends its waters to Laguna Salada. The forms-- There is a series of surface and subsurface forms attributable to the sea. This includes those which owe their origin to the action of wave and current, to the deposition of the alltivium contributed by the river, and to the combined action of river and tide. Of beaches there are three types : mud beaches, sand beaches, rock beaches. The latter two are absent within the delta area proper, being found only marginally. Near the mouth of the river the shore of the gulf is a high, nearly vertical bank of bud. South of this it becomes gently sloping, is irregular in form, and exposes bars and spits at low water. In its major part this mud shore is without relief and, is unbroken by vegetation. Beginning perhaps nine miles north of Point San Felipe is a sand dune barrier parallel to the coast. At high water there is a sandy beach; at low water a mud beach. Inland from the dune is a section of alluvium unmixed with sand, which lies at 'an elevation possibly a little less than sea level, contains driftwood, and shows signs of recent inundation. The dune strip extends northward along the coast nearly to Ometepec Bay, twenty-seven miles north of Punta San Felipe. At a point twelve miles north of San Felipe the sea breaks through the dunes and there is a water-filled channel running parallel to the dune, extending to and entering Ometepec Bay." For the sand of which the dune is composed there is only one possible origin : it has come from the west. How it should tike form as an active dune, lying in part between channel and sea, is not so clear. It could be a matter of uplift and subsequent depression or it might be that the sea had built an off-shore bar upon which the sand lodged. as The dune and dry channel were observed in the field. The information concerning the channel extending south from Ometepec is from a German fisherman, long familiar with the area, whose accuracy of statement has been checked.
]9321 Kniffen: The Natural Landscape of the Colorado Delta 179 Ometepec or Ometepes Bay, if all the variously dated reports are to be accepted, has had a varied history. The Narragansett Survey of 1873-75 reported this bay to be a circular body of water about three miles in diameter, connected with the gulf by a channel three hundred feet wide and one-fourth mile in length. In 1926, Sykes" reported the closing of the entrance channel, while a fisherman of the area is responsible for the statement that at present there are three channels and that there is a bay again. This whole line of the mud beach is hard to examine. It is difficult to approach by sea because of the shallowness and great tidal fluctuation. By land it is equally difficult because the slight elevation of the terrain and the great tidal range mean that each high tide covers a large area and on retreating leaves behind it a liquid mud very difficult to traverse. The only rocky coast in the area is a little stretch above the point of San Felipe. Here at high tide appear rock bluffs against which the waves beat; at low tide there are cobble beaches, sand beaches, and also a beach of larger angular blocks of rubble. The entire point is bounded by a terrace lying about thirty feet above sea level. In part this is composed of the sands and gravels at the mouths of the arroyos which separate the several masses which make up the point; in part of well cemented cobbles which lie on a bench cut into the solid rock. In places this conglomerate has been undermined and lies in great sections on the beach. The area about San Felipe also has sandy beaches, and from this point south sandy beaches dominate the coast. Several minor arroyos enter San Felipe Bay and all these have their mouths closed by sandy barriers. Immediately north of San Felipe there is a sand beach with Colorado alluvium exposed beyond it at low water. This sand beach is backed by a bluff of arkosic material perhaps thirty feet in height, which becomes less marked and passes inland to the north where the expanse of river alluvium becomes wider. River and sea-- Another set of forms is produced by the immediate interaction of sea and river. About the river mouth the shores stand as high banks of the very finest alluvium. About the edges these banks are dissected and gullied by the occasional overflows which cover them. The plains which run back from the channel are without vegetation but have a considerable amount of driftwood cover. 84 Sykes, op. cit., 242.
180 University of California Publications in Geography [Vol.. 5 Gore and Montague Islands are low and flat, and are composed of regular clay strata about one inch in thickness. Their shores rise as steep banks to a height of sixteen to thirty feet at low water. The edges are dissected as a result of the run-off of occasional floodings. There are a few knobs on Montague which are never covered with water. On the surface of the islands there is a driftwood and saltgrass cover. At a point a short distance south of the mouth of the Hardy and the old Colorado channel an east-west line passes marking the dividing point between the spheres of river and sea. North of this line the alluvial plain slopes away from the river to the west, still a part of the general delta profile. South of this line the plain slopes gently toward the sea and river, the normal form built by the sea. Within the zone of influence of both river and sea forms are developed taking their characteristics from both. The inside bank of the river bend becomes something of a tide flat at high water (pl. 22b). If this is a sufficiently extensive area there my be developed a beach on the outside of the bend. This is brought out the more strikingly by the action of waves on the clay in producing little rounded clay balls or pellets. Well within the area where the delta profile maintains itself are a number of overflow channels which pass away from the river toward Laguna Salada to the west. These channels contain water when a flood stage of the river coincides with a high tide so as to produce an over-bank flow. The channels near the river are broad and shallow, with beds composed of hard cracked clay, strikingly in contrast to the loose, crusted alkali material about them (pl. 22c.) South of the area of overflow channels there are other channels, which lead back to the stream, and owe their origin in part to the return tidal overflow, in larger part to the return flow of a portion of the water which escapes the river by meanS of the distributaries higher in its course. In addition, there are the true tidal thoroughfares which dissect lower lying areas subject to the influence of the tide. Where the area has a closed grass cover the channels tend to be narrow and deep ; where there is no vegetation cover, they are broad and shallow.
1932] Kniff en: The Natural Landscape of the Colorado Delta 181 THE SALTON SEA AND LAGUNA SALADA THE SALTON SEA The problem of the origin of the Salton Sink, of the history of the cutting off of a section of the Gulf of California, of its genesis as a desert lying below sea level, is so closely tied up with the history of the delta that a review of the facts bearing on the problem will reveal much concerning the geomorphology of the area with which this paper is primarily concerned. By a scrutiny of the observable facts bearing on the problem it seems possible to suggest a modification of the long accepted theory. The terraces and other indications of a former higher water level are certainly striking features of the area and of prime importance as evidence. About the Salton Sink these features have been long observed and frequently described. Here they have been traced as an almost continuous line. This ancient beach line, as it is called, varies in elevation from twenty-six feet near Centinela to fifty-eight feet east of Brawley." Where the line passes through alluvial deposits or along the sandy floor of the desert it finds expression as a well preserved beach. Where it abuts rock masses it is indicated by their slight notching, and more strikingly, by deposits of calcium carbonate or tufa, appearing only on the more resistant rocks, which are in this case granitic, and lying below the old beach line. But what is more important concerning the terraces is the fact that they do not end, as generally indicated on the maps, but extend on to the south, bounding the whole upper gulf. Instead of ending at Cerro Prieto, the terrace extends along the flank of the Cucopas and Sierra Mayor and swings westward into the Pattie Basin, encircles this, and passes to the south toward San Felipe. On the east the terrace appears as the Yuma Mesa, the Santa Clara Mesa in Sonora, and so on to the south along the eastern margin of the gulf. On the whole, the ancient beach line to the south of the familiarly recognized area is not so well preserved. The river terraces along the Colorado and Gila bound these streams for a great distance above Yuma; they lie at approximately the same height above the river flood plains as does the ancient beach 35 See U. S. Geol. Surv., Reconnaissance Map of the Salton Sink, California (1906).
182 University of California Publications in Geography [vol.. 5
line above sea level. Moreover, they are continuous with the terraces
which form the old beach line, so that it is impossible to say where one ends and the other begins.36 Brown87 suggests the three possible explanations of the origin
of the river terraces as being: (1) diversion of the Colorado into
the dry Salton Basin ; (2) uplift about the head of the gulf; (3)
climatic change. The first he eliminates because of the time factor;
the time required to fill the sink and so reduce the gradient of the
stream below the cutting point not being sufficient to account for all
the material removed at the time of the formation of the terraces.
He rejects the idea of uplift by saying that the water body within
the old beach line was fresh, and can, therefore, not represent an
uplifted portion of the gulf. He is inclined to favor the idea of
climatic change as permitting an excess of evaporation over inflow
into the fresh-water lake and so allowing a sufficient period of time
for the removal of the river alluvium.
The difference in slope of the delta, as between the Salton Sink
to the north and the gulf to the south, has already been mentioned (fig. 2). The steep northern slope, Mendenhal138 says, is to be
accounted for in part by the fact that it represents deposition in the
still waters of the lake which existed there; the flat and relatively
uniform slope toward the gulf representing a stream grade determined
under usual conditions.
Apparently somewhat independent of the true delta slopes is an
elongated depression which borders the delta on the west and is
occupied in part by the New River, the old bed of Volcano Lake,
and the Hardy.
Volcano Lake, extinct since the diversion of the Colorado to the
Pescadero, was long a fluctuating body of water whose volume was
dependent upon its annual increment from the overflow of the Colo-
rado. Accounts as to its former depth are conflicting. It was spoken
of as being "bottomless." In the latter part of the past century
supply boats sailed over it and conveyed provisions to the cow camps
on the Pescadero. At the time the Silsbee map was made it was an
ephemeral feature depending on the annual overflow for its water.
Silsbee" himself, in a deposition concerning Volcano Lake, said that
39 John S. Brown, op. cit., 36. 37 Ibid., 36-40. 38 Op. cit., 20-21. 39 From the records of the Colorado Land Co., Los Angeles
1932] Kniffen: The Natural Landscape of the Colorado Delta 183 it contained water only during high floods, and that at low water there was a channel connecting the Paredones and Hardy which passed through the bed of the lake. To the west was Laguna Prieta, about five acres in extent, which always contained water derived from hot springs. He describes the bed of Volcano Lake as being bounded on the north by a bank but as being low and willow girt on the south. When the water level of the lake reached a sufficient height, part of its waters overflowed through a high channel passing to the west. So nicely was this channel balanced upon the delta crest that part of the diverted waters passed to the south through the Salt Slough, a tributary to the Hardy, a part to the north into the channel of the New. Whatever the former depth of Volcano Lake may have been, by 1907, after it had been receiving a direct flow from the Colorado via the Paredones, its bed was reported as being some ten to fifteen feet above sea level, this after considerable silting must have taken place." During the period of 1916 to 1920, when the lake was receiving the discharge of the Bee--then the main Colorado channel-- its maximum gauge was 42.9 feet." This means that the basin of Volcano Lake is a feature which cannot be accounted for by the normal processes of delta building; that here is a part of an old channel lying far below the level of the ancient beach line. Grunsky" suggests that this may be a remnant of the old overflow channel from the ancient Salton Sea to the gulf. The Hardy has kept the southern portion of this old channel open because it lies in the lowest part of the delta, forming a natural gathering channel for the flood waters of the main stream. The corresponding depression to the north of the delta crest is now occupied by New River, but previous to 1840 it contained only an irregular series of shallow lakes. The dry bed of Volcano Lake is the connecting link in the depression. Volcano Lake is perched directly on the crest of the delta. A stream would not normally flow from end to end along the delta crest, but rather to one side or the other. By a diversion of the alluvium-bearing waters, first to one side of the crest, then to the other, a section of the depression was closed at both ends by an alluvial barrier, forming the bed of Volcano Lake. so C. E. Grunsky, The problem of the lower Colorado River (Typed manuscript in the General Library of the University of California, in form of report to the Secretary of the Interior, dated June 30, 190'7) :9. 41 From the engineering records of the C. R. Land Co., Mexicali, B. C. 42 Op. cit., 9.
184 University of California Publications in Geography [VoL. 5 Another important piece of evidence is offered in the forms created by running water. With the exception of portions of the south delta slope, all the slopes unfailingly reveal that previously created forms are in process of dissection. This applies not only to the delta slopes, in particular those north of the delta crest, but also to the fans and terraces, all forms constructed by running water (pl. 23c). Eliminating climatic change, there remains only one possible explanation: the lowering of the base-level of erosion. This might come about by the disappearance of a water body, say through evaporation ; it might represent a general uplift of the land with respect to sea level. Previous to 1905 there was no great dissection of the north delta slope only because it was not subject to the action of flowing water
in sufficient quantities and concentration. The heavy overflows of 1828, 1840, 1849, 1852, 1859, 1862, 1867, and 1891 were sufficient to cause the presence of considerable water bodies in the Salton Basin." These overflows most certainly did not represent the whole flow of the Colorado and were carried by the then relatively shallow beds of the New and Alamo rivers. New River seems to have come into existence as a recognized channel between 1840 and 1849; the Alamo appears to be somewhat older. The little lakes which lay in the New River depression were not part of the river channel, yet received their water from the stream." In 1899 Cameron Lake, through which New River flowed, occupied a basin which measured a half-mile in length and was twenty feet deep. The lake was rapidly silting up with the material brought in by the New." During the last decades of the nineteenth century there was a minor break in the Colorado near Algodones, occurring annually at the time of the summer flood. A portion of the diverted water went down to the Salton Basin in the channel of the Alamo. A greater portion passed through the Paredones to Volcano Lake and was there divided, the larger part passing south through the Hardy, the smaller northward through the New. At the time of Anza 'a trip across the area in 1774" there was at least one deep arroyo along the delta crest (the Paredones containing water in places at the time of year he traversed the area 48 D. T. MacDougal and others, The Salton Sea (Washington, Carnegie Inst., 1914) :173. **Job Dye passed through the area in 1832 and claimed that New River did not then exist; he furthermore claims to have discovered it in 1849. J. J. Hill, The History of Warner's Ranch and Environs (Los Angeles, 1927) :97. H. G. Hanks, Second report of the State Mineralogist of California, appendix: (1882) :238. He states that New River was formed in 1840.
1932] Kniff en: The Natural Landscape of the Colorado Delta 185 (winter). On his return trip he crossed lower down the delta slope and mentions nothing that could be New River. With the accidental diversion of the whole volume of the Colorado into the Salton Sink in 1905, the situation was radicallly changed. In a period of nine months the New and Alamo cut their channels to an average depth of fifty feet, with a width of 1000 feet, and extended them for a distance of forty-three miles--this by the process of headward erosion" (pl. 24a). Another bit of evidence, possibly connected with the matter of slope dissection, is the fact that there are, in the lower delta, areas which are free from inundation by tide or river overflow. One of these areas is found about the site of La Bomba, another between the old course of the Colorado and the Pescadero, another to the east of the Hardy, south of Volcano Lake. As the agencies which created the forms have undoubtedly been tide and overflow, the fact that they are now above inundation means either that they are higher or that the agencies which created them have decreased in effectiveness. Historical evidence favors the conclusion that for a period of 400 years the Colorado maintained its course from Yuma to the gulf. As Grunsky48 suggests : from Diaz and Alarcon in 1540 to Ives in 1857 there is no intimation that the river had any other course than that which it maintained previous to 1905. Something of a point has been made of the map of Rocque, issued about 1762, which shows the Colorado and Gila as uniting and flowing into the northern end of a body of water, detached from the gulf. If this diversion took place-, it was noted neither by Consag in 1746 nor by Garces in 1771, who found the Colorado very much as it was previous to 1905. If the Colorado had been diverted, say in 1755, it would certainly have been flowing into the Salton Sink at the time of Garces' visit and would have continued until such time as the sink contained a volume of water high enough to overflow into the gulf." The source of the 48 D. P. Barrows, The Colorado Desert, National Geog. Mag., 12:337-351, 1900:343. 46 Diaries and reports of the two Anza expeditions, translated and edited by Professor Herbert Bolton of the University of California. 47 H. T. Cory, op. cit., 1324. 48C. E. Grunsky, op. cit., 6. 49 In this connection, it is interesting to note an item appearing in the diary of Fray Diaz, kept during the trip of Anza in 1774. From the Indians he learned of a river, twenty leguas (about sixty miles) above the mouth of the GiN, which left the Colorado during the flood period and flowed to the west. After a short distance it entered another large stream of very red water; lower in the course these were joined by a small stream of salt water.
186 University of California Publications in Geography [VOR... map is not known and its general character is such that it may he regarded perhaps rather as an inexact copy of other maps than as utilizing the data of some exact, forgotten observer. From Indian artifacts and legends there is another line of evidence. The Collins Valley Indians have legends of great white "birds" which sailed up the waters of what is now the Salton Sink and discharged white men. They tell of the disappearance of water from the sink and of its sudden return. These legends may be good as supporting evidence, but in themselves they are worth very little. The artifacts are more satisfactory. About the old beach line there have been found circles of stones which have been interpreted as being fish traps." These would at least indicate the presence of man within the area during the period when the basin was filled. Even more positive evidence of human occupation is found in the carvings which appear in the tufa of Travertine Point.5' These carvings, which are undoubtedly of Indian origin, are not only carved in the tufa but they have been covered over by deposits of the same material. The first inference to be made from this evidence is that the rock upon which the tufa deposits appear has experienced at least several distinct periods of exposure to the carbonate-bearing waters of the lake. Whether or not this means several distinct fillings of an inclosed lake is a moot question. If this body of water had been subject to great tidal fluctuations could not these carvings have been made at low water ? The presence of the tufa is in itself evidence of fresh or at least brackish water conditions. This is borne out by the great numbers of fresh or brackish water mollusk shells which are found along the old beaches and on the floor of the desert. They are identical in form with living mollusks now found in the occasional permanent streams and springs about the desert margin," many of which are somewhat brackish." In summarizing and synthesizing the evidence regarding the recent history of Salton Sink there appear certain well founded premises from which to proceed to a theory : 55 From a conversation with Mr. Percy Palmer, of BRAWLEY. Si D. T. MACDOUGAL, A Decade of the Salton Sea, Geographical Review, 3:459460 (1917). 52 W. C. Mendenhall, op. oit., 18-19. 53 These mollusks are described and identified by R. E. C. Stearns, Remarks on fossil shells from the Colorado Desert, The Am. Naturalist, 13:141-154 (1879). Some of the specimens were taken from a depth of fifty feet below the surface. Stearns describes them as being fossilized and semi-fossilized and suggests that they might have lived in minor lagoons and ponds as well as in a large lake.
1932] KNIFFen: The Natural Landscape of the Colorado Delta 187 1. Considerations of the problem must not be restricted to the Salton Sink. The extension of the beach line to the south makes it the problem of a larger area. Yet the water occupying the Salton Basin was fresh, while the water of the gulf to the south was certainly not. 2. The river terraces are continuous with those of the beach line and represent depth of cutting equivalent to the height of the old terraces above present sea level. 3. There is only one definite ancient beach line. There are numbers of terraces but these are distinctly of a secondary and subsequent nature, and represent merely a receding water surface. 4. The desiccation of the Salton Sink area is an event geologically very recent. This assertion is borne out by the freshness of the old beach line and by the Indian carvings in the tufa. The ancient lake represents a brief incident, for nowhere along the old beach line did the waves cut deeply into the solid rock. The undissected nature of the north delta slope offers proof of the recency of occupation of the basin by water. 5. There is abundant evidence that there has been, in the area about the head of the gulf, recent and considerable uplift. As proof of this there is the unfailing dissection of fans and terraces, the existence of areas above inundation within the delta, the difference of over thirty feet in elevation as between different sections of the ancient beach line within the Salton Sink. 6. Finally, the delta is inadequate as a dam behind which there could be impounded a lake with a level of even forty feet above the sea. It is not high enough and in addition there was some sort of channel which cut it to a depth near sea level along the line of the Hardy--Volcano Lake--New River depression. Hypotheses-- As an explanation of the sequence of events leading to the forma- tion of the Colorado Desert, Blake's hypothesis has been accepted almost without question. This hypothesis he first postulated in 1853. At that time he was not aware that the old beach line stood above sea level. He suggested the building of a delta barrier by the deposition of the silt of the Colorado, which, perhaps with the assistance of slight uplift, became effective in closing off, or nearly closing off, the area within the Salton Sink. The latter was maintained as a fresh-water body by the northward diversion of the Colorado. Possibly there remained a connecting channel between lake and gulf, to the west of the delta." At a later date, after more careful work had
188 University of California Publications in Geography [VOL. 5 determined the elevation of the terrace about the Salton Sink, Blake postulated the existence of a lake which stood above sea level and discharged its waters into the gulf.55 He still admitted the possible assistance of uplift in making of the delta an effective barrier. Subsequent workers in the area have accepted this view, largely without modification. Barrows" suggested crustal elevation of the earth beneath the central region covered by the delta; Free" ventures the hypothesis that the Salton Sink has very recently become a part of the gulf trough, lying below sea level. A consideration of all the evidence here noted leaves the impression that a modification of Blake's hypothesis is in keeping with the facts : For a brief time during the period of maximum depression of the trough, the gulf stood at the level represented by the old beach line ; it occupiel the Pattie Basin, it stood along the Santa Clara Mesa. The Colorado entered the gulf at a point near Yuma and extended its delta across the gulf toward Cerro Prieto. As the delta advanced westward and became higher (assisted perhaps by general uplift) the deflection of the Colorado to the northward caused the water enclosed by the extreme head of the gulf to become fresh, or at least brackish. Along the western margin of the delta the tide maintained an opening along the line marked by the Hardy, Volcano Lake, and New River depression. The embayment to the north was subject to tidal fluctuations and through the opening the excess waters passed to the gulf. Then came a general uplift of the area and with it the delta became an effective barrier, cutting off the Salton Basin from the gulf. As the river had been flowing toward the north the steeper gradient now lay toward the gulf, and the river was deflected to the south. With the uplift its cutting power was increased and it formed the terraces which now border it. Without increment the waters of the sink gradually disappeared through evaporation, leaving successive minor terraces. During the subsequent period, up to 1905, the Colorado maintained its course to the gulf, grading and building up the southern delta slope to a gradient normal to stream deposition. During all this later period the Salton Basin has received only small amounts of water, which have resulted in nothing more than ephemeral lakes. "Reports of explorations and surveys to ascertain the most practicable and economical route for a railroad from the Mississippi River to the Pacific Ocean made in 1853-4 (Washington, 1856), 5:2:235-240. 55 In D. T. MacDougal, The Salton Sea, op. cit., 3-5. 56 D. P. Barrows, National Geog. Mag., 11:340 (1900). 57 E. E. Free, in MacDougal, The Salton Sea, 25-29.
1932] Kniff en: The Natural Landscape of the Colorado Delta 189 LAGUNA SALADA As previously noted, Pattie Basin is the name applied to the area lying between the Cucopa-Mayor and Juarez ranges. It includes an area perhaps fifty miles long by twenty wide. It is definitely a part of the delta and continues the general delta slopes, having an elevation of about ten feet above sea level at its entrance south of Mayor and sloping down to the northward to a minimum elevation of some five feet below sea level. In the varied history of the water body which has occupied it, in its aspects as a delta feature, there is a strong similarity between this basin and the larger Salton Basin to the north. Both are basins lying below sea level, both are subject to occasional flooding, and both are cut off from permanent occupation by the sea by means of alluvial barriers. The occasional floodings come as a result of the conjunction of high water in the river and high tides in the gulf. This interaction of forces means a backing up of the river waters and a breaking over the banks at a point near the mouth of the Hardy. From this point, several broad, shallow channels run to the westward, passing south of the point of Mayor (pl. 22c). Here the channels become less definite and the water spreads out in a shallow lake. To the northward and close to the flank of Mayor, there is a series of gathering channels, deeply cut into the silt. These are finally joined into one channel which debouches onto the plain about twenty miles northwest of Mayor. North of this is the area occupied by the lake, Laguna Salada or Maquata, during its intermediate and lower stages. When there is sufficient water, the lake extends much farther south. In addition to the historical records, which show the basin to have been occupied by lakes in 1884, 1893, 1906, 1914, 1923, and 1928,55 there is evidence in the form of strand lines. The highest of these is probably a continuation of the ancient beach line, the others are less well marked, frequently being indicated by a line of vegetation. As the lake derives its water from the overflow of the river, it is, of course, fresh at its higher stages. With the loss of volume the water becomes more saline, and finally the lake is reduced to the brine and salt deposits which occupy the northeastern part of the 58 A part of this information as to dates from D. T. MacDougal, The Desert Basins of the Colorado Delta, Bull. Am. Geog. Soc., 39:719, (1907).
190 University of California Publications in Geography [VoL. 5 basin. According to one visitor, who camped at the northern end of the lake in 1923, the water was potable there. What is more striking is his account of the tidal effect, which he says amounted to a movement of six inches, occurring twice a day. This, of course, does not infer a direct connection with the sea, but means rather an alternate increase and decrease in the head of water forcing its way into the basin. Positive proof of the freshness, or at least brackish qualities, of the water of the former lakes, is found in the presence of the great numbers of mollusks which are found about the floor of the basin. Orcutt" stated that these are of the same species as those found in the Salton Sink area. Older accounts of the Pattie Basin reveal very much the same conditions as at present. Garces," in the account of his first trip into the area, mentions a "Rio Amarillo," which is unquestionably to be identified as a channel leading from the river toward the Pattie Basin. While a member of the first Anza expedition, he correctly stated the origin of the ephemeral lakes of the basin. Anza's first expedition crossed the upper end of the basin at a time when the lake had nearly--if not entirely--disappeared. About the old shores I they found windrows of fish, stranded by the retreating water, or I unable to withstand the increasing salinity. A similar experience was encountered by the MacDougal" party of 1907, and by Orcutt"-in 1890. As to the origin of this area as an enclosed basin cut off from the sea, there seems to be no necessity for going beyond the explanation offered for the origin of the Salton Basin. Until a very recent time, geologically, Pattie Basin was part of the gulf. The area was included in the region of general uplift about the head of the gulf. With increased elevation the delta slopes were extended southward, closing the space between Mayor and the coastal range to the south. The arm of the gulf became an enclosed basin, and was desiccated through evaporation. Its subsequent history has been largely concerned with ephemeral lakes, depending for their existence upon overflow water from the river. 6° C. R. Orcutt, A visit to Lake Maquata, West American Naturalist, 7:158164, 1891:159. 60 Bolton, op. cit. 61 Bull. Am. Geog. Soc., 39:720 (1907). az C. R. Orcutt, The Colorado Desert, 10th Report State Mineralogist of California: 914 (1890).
1932] Kniffen: The Natural Landscape of the Colorado Delta 191 THE MODIFYING FACTORS The preceding chapters have considered the nature and origin of the grosser forms of the landscape in the area of the Colorado Delta. It now remains to consider the modifying effects of another group of forces, largely climatic, which are persistent and thus effective in the landscape. Ordinarily the work of the river would appear in this latter group of forces, but in the area under consideration the Colorado is an exotic element which derives its power from another region. In addition to its exterior origin, its importance as a sculpturing agent is such as to class it with the other forces of the first order of magnitude. Climatic Elements-- For the lower delta region there are meteorologic stations only at Calexico and Yuma." These stations serve very well for an understanding of the lowland area of the delta crest but they fail to indicate the modified conditions about the head of the gulf or in the highlands of the desert ranges. At Yuma, a record of the years from 1876 to 1926 reveals for January, the coldest month, a mean temperature of 54.4° F., with a mean minimum and maximum of 42.2° F. and 66.7° F., respectively. For the warmest summer month, July, the corresponding figures are: mean, 91.0° F.; mean minimum, 76.6° F.; mean maximum, 105.4° F.64 55 A sixteen-year record for Calexico differs appreciably from this only in that there appears for July a mean maximum some 12 degrees higher. The absolute maximum and minima are: for Calexico," 117° F. and 21° F., for Yuma, 120° F. and 22° F. These figures mean a hot summer and a mild winter. It is not a frost-free region, however, for the forty-seven year record at Yuma 60 A meteorologic station has been recently established at Mexicali, in Mexico, but as yet the observations do not cover a sufficiently long period to be of any value. 64 In a modified Boppen climatic classification, Russell has designated the Colorado Desert by the symbol "BWhh." This means that in addition to falling into the class of driest, or desert regions, the mean January temperature is above 32° F., and that there are three months or more with mean maximum temperatures averaging 100° F., or over. It. J. Russell, Climates of California, Univ. of Calif. Publ. Geog., 2:77-78, (1926). (See fig. 5.) 66 Annual Meteorological Summary, Yuma, Ariz., 1926. 66 Summary of the Climatological Data for the United States, sec. 13. U: S. Weather Bureau (1921).
192 University of California Publications in Geography [VOL.. 5
reveals only eight years in which the minimum did not drop below 32 degrees. The average frost period comes between December 19 and January 21.67 For agriculture this is significant ; as an aid to mechanical weathering it is insignificant. For mechanical weathering it is of importance that the contrast in temperature, as between day and night, be great. This is a phenomenon generally true of desert areas and particularly true of the Colorado Desert in spring and fall. On one occasion, in the lower delta, a field thermometer registered a minimum of 22° F. during the night. By 10 :30 A.M. it registered 88° F.; this in early December ! By common repute it is always 10 degrees cooler at points about the head of the gulf than at Yuma; San Felipe is said to be muy fresco. It is certain that the modifying influence of the gulf is felt far up toward the delta crest. The not excessive elevation of Centinela (2262 feet) is sufficient to produce an occasional snow cover
. In spite of these seeming exceptions, the actual temperatures as given for Yuma and Calexico serve very well for the whole delta region. The apparent temperature contrasts are frequently sensible rather than actual, and this is a matter of comparative relative humidities. As an average condition, the Colorado Delta is an area of low relative humidity
. In Calexico measurements by the California Development Company" over a three-year period showed the average annual loss through evaporation from a free water surface to be 80.66 inches, or 6.72 feet. Measurements of relative humidity at Yuma are published for three hours of the day : 6 A.M., noon, and 6 p.m." The annual means for these hours are : 60, 27, and 27 respectively. In general there is a close correlation between the hottest period of the year and lowest relative humidity. But what is most striking is that the highest means at 6 A.M. are those for August and September, two of the hot months. The figures for this hour may be supposed to reflect conditions as they are during the night, and there is probably no exaggeration in the stories of dripping eaves during the nights of these months. Certain it is that the camper frequently finds himself in need of waterproof covering. Of course, such condtions raise the sensible temperature and make late summer the most dreaded part of the year.
67 Ibid., 3.
A. T. Dept.
Strahorn and Agr., Bureau
others, Soil of Soils: 11
69 Ibid., 5.
1932] Kniff en: The Natural Landscape of the Colorado Delta 193 The presence of the river is of material effect in raising the relative humidity of the lowland area, immediately adjacent to its channel. It has no effect upon the mesa, though the latter may be washed by the stream." In such cases proximity of hygrophytic and xerophytic vegetation is usual. Such conditions of increased relative humidity find their origin in moisture-bearing winds from the gulf. Throughout the whole area of the lower delta it is in general true that southeast winds dominate during late summer, north winds during fall and winter, west winds in the spring and early summer. For certain localities this generalization is subject to modification because of relief or because of such phenomena as the land-and-sea breeze. In this region the wind is certainly significant as a molder of surface form. Its importance in this respect is not to be measured in such terms as "total windy days" but rather in terms of number and duration of winds of high velocity. Brovvn71 made a detailed study of the wind records for Yuma covering a four-year period, having in mind the dominant direction of the high winds. As a resultant he obtained a northwest wind, and from field observation this appears to apply to the whole lower delta. The dunes certainly show the dominance of a wind acting from this direction. Precipitation-- In Calexico72 the average annual precipitation is 3.10 inches ; in Yuma," 3.42 inches. At both places the rainy season falls in the winter months, with a strong secondary maximum in late summer. The rain falling in the former season is cyclonic in origin, that in the latter is likely to be of the thunderstorm type. The region, then, lies in the transition zone
between the typical cyclonic winter rainfall of California, and the summer convectional rainfall of the gulf coast. As regards their importance as agencies in molding the landscape, precipitation and wind are alike in their effect. The heavy, cloudburst type of rainstorm is proportionally of far greater importance than the gentle rain. The summer rains are more likely to fall in the former class than are the winter rains. In August, 1909, 4.01 inches of rain 70 D. T. MacDougal, Bull. Am. Geog. Soc., 38:4 (1906). 71 J. S. Brown, U. S. Geol. Surv., Water Supply Paper 497:14-16, 28. 72 Ibid., 4. 73 /bid., 5.
194 University of California Publications in Geography [Vol.. 5 fell in Yuma in twenty-four hours.74 Such rains, of course, are rare but they are definitely to be considered in an analysis of the landscape. Truly a desert characteristic is the great variation in amount of precipitation from year to year. As an extreme example there may be cited the Yuma data for 1904 and 1905. In the former year only 1.43 inches fell, while in 1905 the total reached the unprecedented figure of 11.41 inches.75 There is frequently a wide difference in amount of precipitation in the same years at Yuma and Calexico, though their average is practically the same and they lie only about sixty miles apart. Weathering-- Within the area of the delta proper the land forms are primarily an expression of the work of the river. The homogeneous, fine alluvial material has been but little affected by the climatic forces of the area within which it lies. Only the wind and the vegetation serve to some extent as modifying factors. The desert ranges which lie about the margins of the delta are, however, in their origin and subsequent forms, expression of conditions indigenous to the area. In the degradation of these rock masses into their characteristic profiles and forms the dominance of forces particularly significant in a dry climate is recognizable: mechanical weathering, the movement of dry, weathered material down slope, cloudburst rainfall, and wind. Great aridity, abrupt temperature contrasts as between night and day, make for a dominance of mechanical weathering. This condition particularly favors weathering through exfoliation and granular disintegration. Chemical activity is certainly not entirely wanting, as is evidenced by the presence of caliche-cemented deposits, colored concentric bands in boulders, and solution along joint planes. Granites-- In broad outline the sculpture of the granitic masses shows an intricate, regular carving. The slopes are steep, the divides symmetrical. In the desert ranges the skyline is likely to be somewhat serrate. The pattern of dissection resembles nothing so much as a series of intersecting and overlapping inverted "V's" (pl. 24b). In detailed examination there is revealed the very great importance of 74 Ibid., 4. 76 Ibid., 3.
1932] Kniffen: The Natural Landscape of the Colorado Delta 195 jointing. The latter not only exerts a major influence in details of weathering, but also plays a large part in determining the direction and steepness of slope of the denuded rock surface. Accompanying the joint-controlled weathering-out of boulders there sometimes appear examples of what may be termed "socket" weathering, forms which resemble the sockets of ball-and-socket joints (pl. 24c). These are not large, reaching a maximum size of perhaps eighteen inches in diameter. They may appear as concave depressions in boulders or as depressions in protuberances of the bedrock. Forms somewhat similar to these are explained by Bryan as being the work of hisolation and solution. Solution gets a start at a point on the surface poorly protected by the iron-stained crust, common to desert boulders.76 Examples of this type of weathering forms were noted here only in sites of maximum exposure to the sun, though this may be only a coincidence. Occasionally, in the desert ranges, the granites and associated pegmatites, when in contact with finely jointed volcanics or schists, act as ridge builders, the other material falling away in detrital slopes. In no case are these major forms. An examination of a typical, desert granitic range, such as Sierra Mayor, reveals in the characteristic inverted "V" dissection pattern what appears to be a well developed drainage system. There are channels filled with well rounded boulders, tanks, cascades and waterfalls--all generally dry. If the area were to become suddenly humid the disposal of rainwater would be adequately handled. This might be used as an argument for recent and considerable climatic change, but in view of the fact that the volcanic ranges of the area offer no such evidence this conclusion is hardly tenable. What suggests itself is that this pattern represents the normal one developed in such A homogeneous mass by the movement down slope of the dry, weathered material, that this has served as a means of concentrating the occasional desert rains, with the superficial development of normal stream dissection channels. The volcanics- The volcanic lavas and tuffs show a much less symmetrical and homogeneous pattern. In form they range from the basalt-capped mesas to an extremely irregular mass of plugs, cones, cores, and dikes (pl. 25a). As opposed to the granitic masses the volcanics tend to 76 Kirk Bryan Erosion and Sedimentation in the PAPAGO Country, Arizona. Contributionsto 'the geography of the United States, 1922. U. S. Geol. SURV., Bull. 730:19-90, 1923:49.
196 University of California Publications in Geography [Vol... 5 preserve the individuality of the several constituent elements. In detail the material varies from highly resistant, massive lavas to the easily disintegrated tuffs. The more massive thick flows develop widely spaced joints and maintain steep slope angles. The ordinary bedded lava flows maintain in weathering intermediate slopes, talus mantled, whereas the tuffs have gentler slopes. In places the tuffs have almost entirely disappeared, leaving the old volcanic cores rising in cliffy slopes. The volcanic areas as a whole show nothing of the development of the dry stream pattern characteristic of the granite. About some of the more considerable volcanic masses the talus pediments show an extensive dissection by steep-walled, dry water courses. The materials are generally angular, showing little rounding effect along the channels. Perhaps the channels are due to the evident lowering of base-level, possibly, as Bryan suggests for similar conditions, to the effect of occasional cloud bursts. The sediments-- The major exposure of sediments is found in the Tertiary deposits located about the eastern side of the head of the gulf. They are composed of intermittent, thin, sharply dipping, sandstone strata, which appear as low ridges in a mass of gritty clay. The latter is highly weathered with a crumby structure, but for want of a disturbing force frequently retains fairly steep slopes. When disturbed it crumbles, and in so doing exposes numerous gypsum inclusions. From its front along the gulf and for a distance of about five miles to the eastward, this Tertiary formation shows an excellent development of badland topography (pl. 25b). It probably represents an exposure of the same age and topographic development as that described by Mendenhall," lying on the western side of Imperial Valley. Conglomerates appear as prominent exposures but rarely in this area. Where they do appear, subject to the normal process of weathering, their resultant forms simulate those of the granitic rocks. Work of the wind-- The frequent strong winds of the desert are important agents in the transportation of the weathered material. Sand and dust storms are sufficiently characteristic of the area about the head of the gulf so that Hardy speaks of it as the "Smoky Coast."78 77 Op. cit., 25. 78 Op. cit., 324.
1932] Kniff en: The Natural Landscape of the Colorado Delta 197 As a positive factor in the building of forms, the wind is represented by areas of dunes of varying types, and by other sand accumulations. The source of the wind-transported material is very largely found in the weathered debris about the base of the mountains. The Colorado alluvium, through its vegetation cover, or because of its alkali encrusted surface, is ordinarily resistant to removal by the wind. Its surface may be broken by the burrowing of animals, by the action of growing plants, by such cultural agents as wheeled vehicles, perhaps in a measure by weathering. Accumulations of such materials are developed by the wind, but these are secondary in importance to the forms composed of sand. There may be witnessed, across the crusted surface of Pattie Basin, the movement of wind-driven sand, very much after the fashion of drifting snow passing over ice. The sand accumulations are found abutting the mountains or in sheltered spots where the wind eddies. Upon the hard surface exposed to the sweep of the wind, accumulations are found only where some protuberance provides shelter. This function may be fulfilled by a piece of wood, wheel ruts, and only occasionally, on this alkali ground, by a clump of vegetation. The whirlwinds so common to the hot summer days are very effective carriers of material. Quite frequently these winds are of sufficient violence to transport not only dust but sand and even small pebbles. Such winds, though in the aggregate they move considerable amounts of material, are so irregular in their nature as to contribute little to the building of discernible forms of either negative or positive relief. In the area examined no positive example of the abrasion of rock by wind-driven material was encountered, but numerous examples of wind abrasion in mounds of sand and silt were found. Fixed, rather than active, dunes exhibit abrasion, for active dunes move before the wind through the migration of each individual particle. Perhaps with the death of the vegetation cover the wind initiates a cutting of the fixed dune. Through the larger mounds the blast cuts a vertical section which lies in a plane parallel to the direction of the prevailing wind. The forms-- In the foregoing discussion of weathering processes there has been some reference to relief forms, largely those attributable to one set of forces. In the disposal of the weathered material, and
198 University of California Publications in Geography [VOL.. 5 in its integration into land forms, a number of forces operate: gravity, wind, running water, vegetation, and, to some extent, zoogene forces. Though these are the principal forces their effect is modified by such extraneous factors as uplift and depression, and by the influence of the Colorado. The resulting forms, then, are frequently complex in origin. Fans, aprons, pediments-- In the building of these forms three forces are dominant: denudation, running water, and wind. The first two vary in importance with the nature of the mountain mass from which the materials are derived. The detrital material about the base of the lower lying volcanic mountains is composed, in general, of coarse, angular fragments possessing little stratification, its position largely due to its slow movement in a dry state, down slope. The higher lying granitic masses, such as Sierra Mayor, are flanked by deposits of well rounded, stratified gravels. About the higher volcanic masses there is evidence of some water transport of the weathered debris in the characteristic "mine dump" deposits, previously referred to. Alluvial deposits of sufficient convexity to be called cones are lacking. Alluvial fans are well developed along the eastern flank of the Juarez, less so along the Cucopas. As previously suggested, the alluvial deposits along the Cucopas are better described by the term "apron" than by "fan." True fan slopes are recognizable as emerging from the reentrants of this range, but they coalesce, without appreciable break, into the uniform slope maintained by the flanking apron. On the western side of the Cucopas there have developed a number of prominent fans which retain their characteristic profiles out over the river alluvium of Pattie Basin. Along the base of the Peninsular Range the development of fans is much more pronounced, with great boulders, gravel, and sand. But even here is preserved a general uniformity of slope that appears to be due to something more than merely coalescing fans. The alluvial deposits flanking the Peninsular Mountains are deepseated and extensive; those flanking the desert ranges are more superficial. This is a natural condition in view of the greater height and more westerly position of the former range. The deep alluvial sands and gravels of the Salton Sink were cited when speaking of the artesian wells of that area. The general configuration of the floor of Pattie Basin suggests the same condition there, with its
1932] Knifiren: The Natural Landscape of the Colorado Delta 199 maximum depression along the eastern side and its silt-mantled slope rising to the west (fig. 3). Well logs are lacking for the area along the base of the Cucopas, but depths of twenty-eight feet below the surface reveal the gravels dipping so sharply as to preclude the possibility of great extent. The most striking feature of these fan and apron forms is the fact that they unfailingly show dissection as the most recent incident in their evolution. About the flanks of the Cucopas the alluvial deposits are mantled with a cover of palo fierro, palo verde, creosote bush, mesquite, ocotillo, and visnaga. This plant covering speaks for a stability of these surfaces, in fact, reduces them to the category of relict forms. The courses of the ephemeral streams issuing from the mountains are steep-walled arroyos cut into the alluvial sands and gravels (pl. 23c). The easiest way of accounting for these forms is to posulate a recent uplift, and this is in keeping with other evidence. Proof of the recency of the uplift is rendered by the absence of any discernible secondary fan building. The most extensive feature of the material derived from the mountains is the great alluvial plain which extends from the base of the mountains east to the edge of the river silt, and from San Felipe north to the point of Las Pintas. The plain is composed of granite derivatives: sands, gravels, and arkosie conglomerates. Through its gently sloping surface the volcanic masses rise as islands. There is no question that it is water-transported material derived from the distant mountains to the west. In its make-up it resembles closely the mesa which bounds the delta to the east. The origin of both plain and mesa must involve similar events. Quite possibly the alluvium forms only a superficial deposit, for there is exposed by arroyo cutting a very different material, a few feet below the surface, which is of the texture of silt, contains irregular pebble inclusions, and maintains vertical walls on erosion. This all strongly suggests residual material. North of Punta San Felipe the road crosses a series of arroyos, for a distance of perhaps ten miles. They are steep walled, with flat, sandy bottoms. Toward the sea they increase rapidly in depth to a maximum of perhaps thirty feet, with a width of two hundred yards. They are not long, having a maximum length of perhaps five miles. Toward the mountains they become shallower and split into little, sandy washes. Here, again, lowering of base-level through elevation of the land explains the origin of such forms.
200 University of California Publications in Geography [Vol.. 5 Desert pavements-- The term desert pavement or pedre gal is applied to a bare, level surface composed of gravel or small cobbles, embedded in a finer matrix. Such surfaces may be divided into two general types : those composed of alluvial gravels, and those composed of detrital material. In the latter case, if the material is transported at all, it is by the slow movement down slope of dry material, on the face of the retreating mountain base or pediment. The first type may be found as an elevated form. More frequently, as on the mesas, it is an area of negative relief. Its surface is composed of gravels of uniform size darkened by "desert varnish" (pl. 25c). Below the surface is found a hard, clay-like material. The origin of such a form is to be found in the work of the wind, which removes the finer material and causes a gradual lowering of the general surface with a resulting concentration of gravels. The terrace surfaces and the fan slopes, as along the eastern base of the Cucopas, exhibit a similar development. The second type is found particularly well developed about the little volcanic hills north of San Felipe. A detailed examination of one of these forms revealed it as maintaining a slope inclination of ten degrees away from the mountain. There is a perceptible, though not angular break between pavement and mountain slope--the latter averaging about twenty-five degrees (pl. 26a). The pavement surface is composed of angular cobbles showing some desert varnish (pl. 26b). It is only one layer deep, and is embedded in a thin stratum of fine material of like origin. Below the fine material is a stiff sand, sufficiently deep and free of rocks as to be in some measure undermined by animal burrowings. The factors significant in the formation of this land form appear to be : (1) the retreat of the mountain slope through denudation ; (2) the movement of the weathered material down slope; (3) the removal of the finer material by the wind. The dunes-- Of the forms of wind-accumulated material there may be distinguished two general types: active dunes, and fixed dunes. The former have little, if any, cover of vegetation, and are continually changing form and position through movement, under the impulse of the wind, of the sand grains of which they are composed. Should they become stationary for any reason they then fall into the second class of fixed dunes. This latter class also includes the accretion dunes which may be composed not only of sand but also of silt and clay.
3932] Kniffen: The Natural Landscape of the Colorado Delta 201 These forms grow about a nucleus initially fixed, the material being held in place by the roots of vegetation. There are two major sets of active dunes, though both are marginal to the lower delta region. The first is the Algodones Sand Hills, lying in the eastern part of Imperial County; the second, the belt of dunes lying parallel to and east of the edge of the Santa Clara Mesa, in Sonora. These latter are not easily accessible and are little known, though they may be readily seen from the edge of the mesa, twenty-five miles distant. It appears that they are at least in part of barchan form and that the dominant wind is southerly. The Algodones Sand Hills have a northwest-southeast orientation, forming what is in reality a continuous range some forty miles long with a maximum width of about five miles. The hills lie upon the smooth plain of the mesa which bounds the delta at the north and east, spilling over a bit at its southern end onto the alluvium of the delta. The irregular forms of the dunes suggest that they are not under the domination of any particular wind. The range is surrounded by extensive areas of desert pavement and it seems possible that in part the dunes owe their origin to the concentration of the material removed by the wind in the formation of the pavements. On the basis of comparison of present form and position with that represented on old maps, Brown" comes to the conclusion that the Sand Hills in their larger features have changed very little, at least in the last seventy years. He suggests a slight movement to the southeast; this is borne out by the presence of extensive areas of fixed dunes, south of the major range, on the delta. Here the conditions of abundant moisture and plant growth are much less favorable to the existence of live dunes. Barchan dunes were observed on the Sonora Mesa near the head of the gulf, where conditions are considered ideal for their development--a hard floor, and a strong wind blowing in a constant direction (pl. 26c). The dunes observed along the coast north of Punta San Felipe are apparently subject to the effects of winds blowing from different directions (pl. 27a). Evidence of the very considerable height to which wind-driven sand may be forced up steep slopes is offered by several examples. The northern tip of Las Pintas, which rises to a height of at least five hundred feet above the alluvial plain, stands directly in the path of the strong winds which sweep south out of the Pattie Basin. At 79 U. S. Geol. Bury., Water Supply Paper 497:29.
202 University of California Publications in Geography [VOL 5 one point, where the slope of the range measures about thirty-five degrees, the sand has been driven completely over the crest and lies as a white band down both sides. Though this is the extreme case, many other examples of less magnitude are observable. In no case was there observed an active dune which the constituent material was finer than sand. Fixed dunes, however, may be composed in part or in whole of materials of the texture of silt or clay. These latter are of the type termed "accretion" dunes. The distinction between active dunes which have become fixed by vegetation and accretion dunes composed of sand is rather difficult. In time both come to approach the same form. The action of roots in destroying the structure, the accretion of finer materials, and the decay of vegetation tend to make the origin obscure (pl. 27b). Good examples of accretion dunes are the mounds which collect around springs owing to the retentive powers of the vegetation which the presence of water induces. One such is the Kane Spring mound in Imperial Valley." There are traditions in the lower delta of an "Agua Hechicera" (witch water) in which the spring emerged from the top of a dune. Another type of accretion dune is that previously mentioned, where the sand, in sweeping over a bare surface, is lodged in a depression or behind some obstruction. Examples of this are seen on the bare, alkali surface of Pattie Basin, where wheel ruts catch the drifting sand. Vegetation takes root in the sandy spot, particularly sea purslane. A series of mounds is built up ; they in turn offer further obstruction to the sand; other species of vegetation are introduced (such as creosote bush), and the accumulation continues (pl. 27c). An examination of a number of forms thought to be of this origin revealed a series of dunes, irregular in form, some with a crater-like top covered with mesquite. On the flanks of some were small active dunes. An examination of the structure revealed a rough bedding with alternate layers of fine sand, magnetite, coarse sand, fine sand, with inclusions of small, irregular rock fragments. Mounds composed of silt or clay material are found in all sizes, from those of a height of ten feet or more to those of no more than a foot. These forms are generally quite smooth as to outline, though they vary in form from those having steep sides to those which are low lying and have gentle slopes. There seems little doubt that these are wind accumulations, held together by vegetation. Even 80 Cited by J. S. Brown, op. CIT., 31.
1932] Kniffen: The Natural Landscape of the Colorado Delta 203 where these have become denuded of vegetation, examination reveals numerous remains of roots. Such forms generally occur in groups and are highly restricted in distribution. An examination of one of these revealed a surface very much hardened by carbonate concentrations. Below appeared the alluvium, soft, moist, cool, and containing plant remains, small black seeds, and small bones. As this particular form lay in the entrance from the Colorado to Laguna Salada it is possible that some factors other than the wind and. vegetation have been significant. Caliche- Caliche is widespread in these arid areas and is frequently known locally as hardpan or cement. It is thought to be derived from ascending ground water, or to be a deposit from rain water, which characteristically effect a solution and concentration of soluble salts in desert areas. In the delta area caliche is found in the pure form exposed along cuts, but it is more significant as a cementer of sands. Along the edge of the Santa Clara Mesa there appear exposures of calichecemented sandstone, fallen and broken into slabs. The sand appears to have been cemented a short distance below the surface, to have been exposed by denudation, then caved and broken by undermining (pl. 28a). It seems quite possible that many of the conglomerates present along the stream channels of this region represent the effect of caliche. SUMMARY OF GEOMORPHOLOGY Diastrophism and the Colorado River are the two agencies of major importance in the genesis of the delta landscape. Diastrophism is significant in that it has determined the potentials of deposition. It also sets certain areal limits within which the river may work By this restriction it has effected the gross configuration of the alluvial surface. The river is significant primarily as a transporting force which has carried into the area a great burden of silty alluvium. In the Delta Cone area the river has affected the detailed relief, almost to the total exclusion of other forces. In a manner, diastrophism and the river are forces which work against each other in the net result on the landscape. Diastrophism
204 University of California Publications in Geography [vox.. 5 has produced relief ; the river has tended to destroy it. In detail the river does produce relief forms. Still, the four major areas of the delta surface mark depressions in the diastrophic framework. Volcanism and recent faulting have produced modifications in the frame. The sea and the climatic forces work with the river to produce a general reduction of relief. In a consideration of the genesis of the delta landscape, as of any other, it is necessary to think of a mutual interaction of forces, going on contantly. Genetically "pure" forms are rare. The nearest approach to such forms is to be found in areas of recent and contemporary volcanic action. The diastrophic frown has not been proved that the depression occupied by the Gulf of California is a simple graben. It is impossible to point out two continuous fault planes between which there is a dropped block of the earth's crust. So far as definite evidence goes the form may as well be a geosynclinal one. The trough does, however, form a zone in which faulting, and to a lesser extent, volcanism, are highly significant in their morphologic implications. At least three phases of diastrophic activity can be recognized in the delta landscape: 1. The forming of the gulf trough. This may represent folding or faulting on a large scale. The trough is a structural form which extends from Coachella Valley to Cape San Lucas. 2. Changes in elevation over a broad area, characteristic of the post-Pliocene period, along the coast of Southern and Baja California. In the delta area the ancient beach line marks a uniform degree of uplift over a large area. 3. Faulting of orogenic and lesser magnitude. It has produced as relief features such forms as the Cucopas, and Pattie Basin. Tertiary sediments have been exposed to the effects of erosion and denudation. To some extent faulting has exerted a control over erosion in the development of surface forms. Volcanism is significant in the delta landscape principally as a producer of relief forms. It has been active in periods so recent geologically (late Tertiary and early Quaternary) that the volcanic forms have been but little reduced by the attack of denudation and erosion. In the case of the Sierra Pintas, volcanism has affected the areal configuration of the delta surface in nearly making of Pattie Basin a mountain-encircled depression.
1932] Kniff en: The Natural Landscape of the Colorado Delta 205 The covering-- In the process of covering the diastrophic frame the function of the Colorado has been to deposit its alluvial burden in the depressions; denudation and erosion have been of major effectiveness in attacking and leveling the eminences. Of the two sets of forces the river makes by far the more impressive showing in the present landscape. Choked by a burden of river silt the sea has ceased to be an active erosive force and has become the agent which, by wave and tide, disposes of the alluvium in characteristic forms. The building of the delta represents only the most recent development in a long continued filling of the gulf trough. As a surface form the delta is no older geologically than late Pleistocene. Alluvium of the same textural character as that now being deposited has a depth on the delta crest of no more than two hundred feet. At the present rate of increment all this silty alluvium might have been carried into the area within a period measured in hundreds, or at most a few thousands of years. Probably the whole delta surface is underlain by Tertiary and Quaternary sediments. At least a part of the Quaternary deposits represents a contribution by the Colorado River of that time. A Quaternary sea stood over the mesas and gave the alluvial sands and gravels their present character as flat-lying, undissected plains. A late Quaternary uplift resulted in the withdrawal of the sea from the mesas. The Colorado was rejuvenated; it deepened its bed, and deposited gravel in the gulf trough. Succeeding this period came a depression in which the Colorado silted up its previously deepened channel. The sea advanced to a point near Yuma, forming what is now recognized as the ancient beach line. With carrying power reduced, the lower river transported only the finest alluvium. From this time, somewhere in late Pleistocene the development of the present delta has taken place. At a time, perhaps 500 to 1000 years ago, the delta emerged as a surface form and became the barrier which separated the Salton Sink from the sea. In effecting the closure between the two water bodies the normal delta building processes were assisted by a gradual and extensive uplift. In the subsequent extension of the river deposits there have been formed three types of surface, each characteristic of the manner of deposition of the material. These three types of surface are exemplified in the four natural areas into which the delta may be divided: the Delta Cone, Pattie Basin and the Salton Sink, and the Gulf Plain.
206 University of California. Publications in Geography [VoL. 5 The Delta Cone possesses a gently convex surface, formed by the pendulum-like swing of the main river channel. The river channel becomes relatively high ground through deposition, forcing the stream to seek a more favorable gradient. Since the cutting off of the Salton Sea from the gulf, the activity of the river has been mainly restricted to the area lying south of the delta crest. The area north of the crest has been served by overflows from the main channel; occasionally the overflows have caused the diversion of considerable amounts of water. Laguna Salada and the Salton Sink have the concave profiles characteristic of deposition in still water. In their previous condition as estuaries of the gulf these areas undoubtedly received alluvium derived from the Colorado. At present they are dependent for any additional deposition of silt upon the occasional flood waters from the river. At such times when they are flooded, both are the sites of lakes in which the silt is precipitated. In cross-section, the Gulf Plain resembles a very flat "V." This configuration has been effected by the tide in the manner characteristic of estuarine deposition. The tide advances with considerable force over the flat, carrying the silt with it. The ebb is marked by a less concerted retreat of the water, which remains in shallow pools, or trickles back slowly; silt is deposited. The climatic factors have been of almost no significance in the development of the delta surface. With the exception of the dunes and occasional patches of alkali, they find little direct expression. The alluvium remains essentially in the same state and position as it was when deposited. The soil horizons--if they may be called such--are horizons of deposition rather than of alteration. The climatic forces find their major activity in an incessant attack on the mountain masses. At the southern end of Pattie Basin there are several little granitic hills, nearly submerged in a plain of sand and gravel (pl. 30a). These forms reveal the near victory of desert aggradation; the steep escarpment of the Sierra Juarez reveals the opposite case. Mechanical weathering is more significant than chemical weathering because of the nature of the climatic factors: scanty precipitation, and sudden and considerable temperature contrasts. Denudation is more significant than erosion, partly for the same reasons--though the occasional occurrence of cloudburst rainfall makes erosion a factor to be considered.
]932] Kniffen: The Natural Landscape of the Colorado Delta 207 The residual surface of the granitic masses possesses a pattern which is regularly and quite intricately developed, since the rock is regularly jointed and of homogeneous composition. The volcanic masses are irregularly jointed and heterogeneous in lithologic composition. As a result their residual surfaces are highly irregular in form. Along the base of the Cucopas there are considerable colluvial deposits, unfailingly dissected, which give the impression of being relict. As these deposits exhibit beveling by the sea at elevations approximating those of the ancient beach line, they can be referred in age to the time when the Salton Sink was a part of the gulf (perhaps a thousand years ago). It would be venturesome to say that alluvial fans are not forming at present, but certainly the building of these has not proceeded far enough to make them readily discernible. Somewhat the same case holds for the volcanic ranges. Here there are sandy washes which appear occasionally to contain water. There are forms, constructed of colluvial materials, certainly relict. One such form, particularly striking, is found at the base of the Sierra Pintas. It is flat topped, perhaps three hundred feet long, and rises to an elevation of twenty feet above the tide flat by which it is surrounded. The form is composed of sand and gravel, with enough finer material to retain fairly steep sides. It is separated from the apron of colluvial materials which fronts the mountain, by a distance of about a hundred feet. Such a form is a marker of uplift, but it also indicates a former activity in transportation and deposition of weathered materials that is certainly not apparent at present. For the whole delta area probably the most troublesome problem is concerned with the matter of colluvial deposits. Briefly summarized, the significant points are as follows: 1. Along the base, both of the Cucopas and of the Sierra Pintas, the colluvial deposits possess forms which are characteristically flat-topped, terminating outwardly in an abrupt face. In both cases the form: appear to be relict, are mantled by trees, and are unfailingly dissected. The Cucopas deposits are composed of well rounded and partly stratified materials; the Sierra. Pintas deposits are composed of irregular fragments, not appreciably stratified. 2. The only fans possessing a normal profile are found along the western side of the Cucopas. Even they are heavily mantled with vegetation. 3. The building of secondary, or new fans, is not discernible. In an explanation of the forms--or lack of forms--uplift is a factor admitted without question. For the rest there appear to be two possibilities: simply, climatic change or no climatic change.
208 University of California Publications in Geography [VoL. 5 Climatic change of such a nature that fans are no longer formed, solves the problem immediately. With stability of climatic conditions it is necessary to introduce a time factor in which uplift is involved: the process of building the forms is such a slow one that sufficient time has not elapsed since the uplift to permit of the formation of subsequent deposits. Particularly about the margins of the Delta Cone is the contact of river alluvium and colluvial materials sharp. The line of distinction is least clear along the base of the Sierra Juarez, where the occasional mountain torrents have carried sand out onto the surface of Laguna Salada. The wind has also contributed to a mixing of sand with the silt. Its activities have been least significant in that portion of the Delta Cone possessing a heavy plant cover. The distribution of dunes bears a relationship both to the source of the sand and to the direction of the strong winds. The two major dune areas (Algodones Sand Hills and Sonora Hills) are both located on the mesas, at least twenty-five miles from their western margins. The large areas of desert pavement on both mesas suggest a source of the sand. At least for the Algodones area the dominant, strong winds blow from the west and northwest. In Pattie Basin the dunes show the dominance of the north wind. The source of the material is to be found in the colluvial deposits. So strongly does the wind sweep through the basin that the dunes are not found out on the alluvial surface, but only in places where irregularities in the mountain front offer protection. In the sheltered spots the dunes rapidly take on a growth of vegetation and becomes fixed. THE NATURAL AREAS On the basis of a systematic combination of morphologic features five natural areas are distinguishable in the Colorado Delta: the Salton Sink, the Delta Cone, Pattie Basin, the Gulf Plain, and the desert mountain areas, the Cucopas and Las Pintas ranges (map 1). The Salton Sink is sufficiently characterized in literature and is here disregarded. The Delta Cone-- This area is so named and so distinguished because it is only this area of the five which possesses the profile and surface configuration generally associated with the term "delta." The other areas are included in the delta region only by virtue of their alluvial sur-
1932] Kniff en: The Natural Landscape of the Colorado Delta 209 faces, or because they occupy contiguous positions. The Delta Cone is the heart of the delta, and in it are presented most vividly the striking contrast between the well-watered channel of the Colorado and the arid surroundings through which it flows. The surface of this area forms a broad, flat cone, with gentle slopes averaging less than two feet to the mile. The northern boundary of the area is placed at the line where the degree of slope increases sharply toward the Salton Sea. To the south the edge of the cone meets the Gulf Plain at a point where the river gives way to the sea as a dominant force in the landscape. The area is really a great plain, broken by no forms possessing considerable relief, but with a surface minutely broken by characteristically deltaic features: meandering channels, lakes, natural levees, and the somewhat specialized forms of the secondary deltas with their anastomosing channels. This area, then, in its surface features, represents primarily the work of the river. Even the Hardy, Volcano Lake, and New River depression, which are remnant portions of the old tidal thoroughfare connecting the gulf and the Salton Basin, have been greatly modified by the activities of the Colorado. In addition to these remnant forms there are a few others whose origin cannot be traced to the river: Mesa Andrade, the mud volcanoes, and the dunes. The dunes are found extensively only in that part of the region which was formerly poorly covered with vegetation. This includes most of the area lying north of Volcano Lake, and this is the area most infrequently flooded by the Colorado. The area was freely open to the northwest winds, which formed dunes with sand derived from the mesas bounding the Delta Cone on the west. In the present cultural landscape, a portion of these dunes has been leveled, and the others have become fixed by vegetation. That the dunes were formerly numerous and extensive is ably borne out by the accounts of early travelers, who found great difficulty in forcing a passage through the area east of Centinela." The climatic factors have been of little significance in the development of the morphology of the Delta Cone. The river alluvium is already in a state of such fine subdivision as to defy further disintegration through mechanical weathering. In spite of high temperatures and abundant water, even chemical weathering is not highly significant, possibly because of the scarcity of hurnic acids in the
210 University of California Publications in Geography [voL. 5 soil. Chemical activity is represented by alkali concentrations, and by slight inclusions of organic material in the upper soil horizon. The soils of the Delta Cone, and in fact of the whole alluvial surface, represent deposition rather than residual development. Their case is analogous to that of fresh glacial or loess soils. In 'texture they vary between narrow limits, from an adobe to a very fine sand. In their areal distribution over the surface of the delta they are indicative of the conditions under which they are deposited. The sandy soils are found close to the stream channels; the adobes are found in the beds of lakes and lagoons.82 The agriculturist of this area makes soil distinctions based largely on texture. He recognizes the soil horizons as being depositional forms, and frequently can name the area from which the alluvium was derived. One advanced the following classification of soils : 1. Gray alluvium which cracks and remains coarse. (Probably a silt.) 2. Gray alluvium which breaks up into rather fine particles. (Silt mixed with small amount of sand.) 3. A reddish, adobe-like alluvium, as in the bed of Volcano Lake. (Colloidal material.) 4. A more broken, reddish alluvium. (Silt?) 5. A fine sand or quicksand. 6. A mixture of all. Another advanced the following: 1. Hard, with something resembling hardpan. (Silt or adobe with a subsurface concentration of soluble salts.) 2. A "buckshot" soil. Is the best agricultural soil and retains water well. (Mixture of silt and sand.) 3. Sand. Does not retain water. These practical classifications have seized very well upon the distinguishing characteristics of these soils. It is notable that they make no place for alkali soils, and that is quite fitting, for on the wellwashed and well-drained slopes of the Delta Cone there are no considerable alkali concentrations. An areal classification of these soils becomes a matter of tracing out the recent morphologic history; with the knowledge of the peregrinations of the river it is possible to predict the type of soil to be found in any particular area. The alluvial burden of the river is so great and the deposits so thick that it is possible that there should be a complete change in soils with the flood of one summer. 82 An excellent statement of these soil conditions is found in: A. T. Strahorn and others, Soil Survey
of the El Centro area, California, U. S. D. A., Bureau of Soils, 18 (1922).
1932] Kniffen: The Natural Landscape of the Colorado Delta 211 A parallel case is that of the vegetation, which also shifts and changes with the course of the stream, and, to some extent, is associated with changes in the upper soil horizon. Within the mesophytic formation of the delta there is a nice adaptation to conditions of soil and water. As these conditions are changed by shifts in the course of the stream, so is the balance disturbed. Where conditions are long stable, as they were about the old channel of the Colorado, and as they are about the head of the gulf, there is a corresponding adaptation and stabilization in the plant world. The screwbean (Prosopis pubescens) and mesquite ( Prosopis velutinea) are rather widely distributed, and are found in areas of stable soil and water conditions. The cottonwood (Populus maedougalii) is dominant in strips which border the old channel of the Colorado (pl. 28b). Saline soils introduce special conditions which reduce the number of possible plant occupant. Characteristic of the saline soils about Volcano Lake are Achronychia cooperi and Atriplex fascidulata. In the areas subject to occasional tidal overflow are found salt grass ( Distichlis) and Cressa (pl. 30b). The mesophytic plant formation of the Delta Cone is striking evidence of the modifications introduced into the desert by the presence of a large stream. Nor is this entirely a matter of increased water supply. Edaphic modifications wrought by the river are of decided importance, so that the line of contact between the river alluvium and the sands and gravels of the colluvial aprons becomes a significant one. There are, about the margin of the delta, alluvial areas that are unwatered or alkaline. In such cases the vegetation is wanting or effects specialized forms to meet the conditions. But frequently the marginal contrast is a startling one, with an abrupt transition from mesophytic to xerophytic plant associations. On the silty floor there may be a dense growth of mesquite and arrowweed, while within a short horizontal distance the vegetation of the mesa is dominated by ocotillo, creosote bush, and cactaceae. As the Delta Cone has a unique and vigorous vegetation cover, it is the home of a varied and abundant animal life. In the streams are found the beaver, muskrat, and a variety of fish. Land animals range in size from the mule deer and "lion" through the coyote to raccoons, rabbits, and lizards. Game birds include both aquatic and upland forms, while the non-game birds are represented by the crow, blackbird, turkey buzzard, and numerous other species.
212 University of California Publications in Geography [Vca.. 5 Pattie Basin-It is quite possible that Pattie Basin represents a tectonic feature comparable to Borrego Valley, north of the international boundary. It is bounded on the east and west by granitic ranges; along the base of each there runs a fault line. Definite evidence is lacking, but from the surface configuration of the basin it seems probable that it contains a deep fill. Colorado alluvium forms a surface veneer which slopes gently from the base of the Cucopas upward to the base of the Sierra Juarez. Underlying most of the surface is a substratum of sand and gravel derived from the Sierra Juarez. The alluvium is a lacustrine deposit, undissected by stream channels, and bounded by a series of strand lines. The highest of these is the most marked and probably represents a continuation of the ancient beach line (pl. 23a). The others are less well marked and represent stages of the ephemeral lakes that have occupied the basin from time to time. In part they are distinguishable only by virtue of the line of vegetation which marks them. Within the basin several surface types are distinguishable. The alluvium is composed of colloidal material, but around the edges there has been something of a mixture with wind-blown sand. The basin is without drainage so that there has been extensive concentration of soluble salts. The vegetation cover is in no case abundant, but on the saline surfaces it is entirely lacking. On a basis of soil texture, salinity, and vegetation cover, the types of surface may be distinguished as follows: 1. The salt beds. These lie in the lowest part of the depression when it is not occupied by a lake. Extensive deposits of crystallized salt of commercial importance. 2. The black-alkali surface. A smooth surface, without covering of vegetation. Moist and soft, beneath the surface, the moisture being held by the calcium and magnesium chlorides in the soil (pl. 28e). 3. The baked, crusty surface splotched with small hummocks which break under foot. Nearly barren of vegetation. The hummocks are characteristic developments on a drying surface (pl. 29a). 4. Shallow pans or troughs with a surface of hard,, cracked alluvium. Little or no vegetation cover. 5. Flat surface of mixed sand and alluvium. Small accretion mounds of the same material capped with Segavium. 6. The sand hills. Fixed by vegetation. With flat tops which preserve something of a general level. 7. Colluvial sands and gravels. The apron which lies at the base of the Sierra Juarez. Composed of sand, gravel, and boulders. Cut by sandy washes. Abundant vegetation cover dominated by palo fierro and mesquite.
1932] Kniffen: The Natural Landscape of the Colorado Delta 213 In the main the alluvium forms a great barren plain, with extensive areas perfectly destitute of vegetation, marked only by the white salt deposits or by patches of moist alkali soil. About the margin of the alluvium are a number of water holes. With these there is introduced a comparative abundance of life-forms. Pozo Cenizo lies near the edge of the gravel terrace north of the tip of Las Tinajas Range, and marks the site where wells are dug down to a depth of eight to ten feet to water. The presence of water here probably represents nothing more than a subsurface sand or gravel stringer which reaches out from the terrace. About this site there are several mesquite trees. Fairly dense patches of arrowweeds are bordered by salt grass. Animal life ranges from the occasional wild burro and coyote to the desert mouse. Pozo Coyote, at the southern tip of Sierra Mayor, is another well of the same type as Pozo Cenizo. Its site has something of a parklike appearance, with mesquites rising out of a mat of salt grass. In common with the other wells of this area Pozo Coyote must be dug out and bailed before the water is usable (pl. 30c). Farther north along the base of the Cucopas there is a hot spring whose waters register a temperature of from 112 to 128 degrees F. This lies along the fault line which bounds Pattie Basin on the east. Coming into the area after the recent desiccation of one of the ephemeral lakes, MacDougal recognized a number of strand lines in terms of a banding of vegetation. For the highest strand he noted a zone of mesquite, below it a narrower zone occupied by salt-bush and mallows, while the lowest bore only clumps of sea purslane.88 At the present time the surface has been uncovered for a period sufficiently long that the purslane extends far out over the plain, growing in scattered sandy hillocks. The Gulf Plain-- This is the term applied to the area about the head of the gulf, where the littoral deposits are formed of alluvium derived from the Colorado. This is the realm of the sea, in which tide and wave have constructed a characteristic set of forms. The area occupies a portion of the diastrophic depression, the gulf trough. It is bounded on the east by the San Jacinto fault and the Santa Clara Mesa, on the West by the Sierra de las Pintas. 83 D. T. MacDougal, The Desert Basins of the Colorado Delta, Bull. Am. Geog. Soc., 39:722 (1907).
214 University of California Publications in Geography [VoL. 5 The alluvial fill in this depression must be a considerable one, for the water shoals rapidly from a depth of eighty fathoms east of San Felipe to a depth of eight fathoms east of Otnetepec Bay. There is no reason for believing that this represents a rise in the rock floor which underlies the alluvium. An interesting, but at present unexplainable, feature is the granitic island, Consag Rock. It has a pillar-like form and is surrounded by deep water. The natural inference is that it represents a fault sliver, but its structural affiliations are uncertain. In profile the Gulf Plain resembles a flat "V," with the shores sloping gently up from the waters of the gulf which occupy the apex. This is the normal surface configuration to be expected of a sea-built shore; it stands in striking contrast to the gently convex surface of the Delta Cone. Where the energy of the tide is concentrated, as near the mouth of the river, the shore line of the Gulf Plain is a nearly vertical bluff. Farther south, where the tide is of less influence as an erosive agency, the shore line is gently shelving. Instead of a bluff it exhibits such features as bars, spits, and lagoons. The surface of the Gulf Plain is generally barren and without relief. When wet this surface is composed of a very sticky mud. When dry it is salt encrusted. A portion of it, about the outer margins, is not subject to inundation even by the highest spring tides. This condition has probably come about through uplift. Along the shore there is a considerable cover of driftwood, brought down by the Colorado. In great part the plain is without vegetation but there are occasional clumps of alkali weed ( Cressa) to which the mirage gives weird properties. Near the mouth of the river the plain is broken by the little channels which are tributary to the gulf. Their course across the plain is frequently marked by a border of vegetation; mesquite and .Atriplex. The association of vegetation with these channels is not so much a matter of the occasional presence of fresh water as it is of reduced salinity of the soil. The boundary between Delta Cone and Gulf Plain is one which has shifted southward with the progressive extension of the river channel. It is a slow process which has been carried out by innumerable floods and meanderings of channels, with the superimposing of the convex surface of the Delta Cone over the Gulf Plain. It has meant an extension of all the forms which characterize the Cone na ural leve s 1 oons, and dense ve etation.
] 1932 Kniff en: The Natural Landscape of the Colorado Delta 215 Since the slopes of the two surfaces are gentle it is difficult to distinguish their line of contact by reference to profile alone. Much better as a marker is the contrast in vegetation introduced by differences in salinity of soil and water. Cottonwood and willow form a characteristic border along the old channel of the Colorado, but they do not grow on ground moistened with brackish water. In 1907 MacDougal found that the tide ascended to a point near Mesa Andrade and that the cottonwood-willow association did not descend beyond that point. Yet within the memories of old Indians then living, the limit of those trees had advanced a distance of eight or ten miles southward.84 In advance of the cottonwood-willow association, on land newly claimed by the Delta Cone, are found the mesquite and Atriplex. These plants are somewhat more tolerant of saline conditions. Tules grow even along the banks of streams affected by the tide, and sloughs of "wild rice" ( Uniola palmeri) are found. On the saline soils are found the halophytes: iodine bush (Spirostachys occidentalis); on the mud flats, sedge (Scirpus fluviatilis) ; on the land subject to tidal overflow, salt grass and alkali weed (pl. 30b). To compensate for this loss at the north, the Gulf Plain has experienced steady growth at the south. Sykes' map of 1907 shows the plain as reaching to a point about ten miles south of Ometepec Bay. At the present time (1928) it extends, in an ever narrowing wedge, to the rocky mass of Punta San Felipe. Through a combination of tide, wave, and current, there is first effected a shallowing of the water along the shore. In the next stage the alluvium appears as an exposure at low tide (pl. 27a). In the final stage it appears as a wide plain, inundated only by the highest tides. Were the old regimen of the Colorado continued, the Gulf Plain and the Delta Cone would advance to the south in the gulf trough. Upstream from the Delta Cone there would be developed the normal valley profile of the river having great seasonal fluctuation in volume : natural levees, with the flood plain sloping gently down to the valley walls. As it is, with the diversion of much of the river's volume, and the loss of much of its alluvial burden, it is quite probable that there has been reached something of a condition of stability, and that further shifts will be slight. as D. T. MacDougal, Botanical features of North American deserts (Washington, Carnegie Inst., 1908) :34.
216 University of California Publications in Geography [Vol.. 5 The contact between the alluvium of the Gulf Plain and the bordering sands and gravels is a distinct one. Since the effect of the ephemeral streams is slight, the only mixture of silty alluvium with sand is through the medium of the wind, and this is not considerable. Equally abrupt is the transition in vegetation cover, from the nearly vegetationless Gulf Plain to the sands and gravels, with their highly specialized xerophytic flora (pl. 29b). The desert ranges-- THE CUCOPAS This natural area includes the mountain chain extending from Centinela on the north to Sierra Mayor on the south, and in addition the flanking deposits of sand and gravel. It escapes being an island in a sea of silt only by virtue of the low rocky ridge which connects it on the north with the Peninsular Range. Though it forms a topographic unit and may properly be considered a natural area, the range is not at all homogeneous in make-up nor is it a simple and unit structural feature. From north to south along the range there are recognizable four distinct structures: (1) the semi-isolated Cerro Centinela ; (2) the granitic massif of which Cerro Borrego is the culminating peak ; (3) the metamorphic and volcanic hills, locally known as the "Cucopas;" (4) the granitic Sierra Mayor. With the exception of the "Cucopas" the whole range is essentially granitic, but there appear patches and exposures of other rocks which are suggestive concerning the geologic history of the area. A dark diorite flanks Sierra Mayor in places along its eastern side and appears in patches on some of its higher peaks (pls. 19e, 20c). The contact between the two is sharp, the diorite being somewhat metamorphosed. Farther north along the base of Mayor the diorite is succeeded by a highly distorted schist, and at one site there is an exposure of a bedded quartzite, recently faulted. The presence of these distorted, metamorphosed rocks may be indicative of the laccolithic nature of Sierra Mayor. The low-lying hills north of Mayor possess a granitic core, intrusive in a mass of highly distorted and broken metaraorphics. Near the contact of these hills with Sierra Mayor there is found a deposit of sulphur, probably derived from recent solfataric activity. Flanking Centinela on the north and west is a massive sandstone, probably Tertiary in age. High on the side of the mountain this same formation appears as a true quartzite.
1932] Kniff en: The Natural Landscape of the Colorado Delta 217 Minor faults are abundant and distinct, and the evidence is sufficient to postulate a more extensive one which continues throughout the length of the range. It lies along the western sides of Centinela and of the Cerro Borrego massif. Here its course is marked by an escarpment and hot springs. It passes through the Cucopa Hills and along the eastern side of Sierra Mayor. In the Cucopa Hills the course of the fault is at least suggested by the sulphur deposits, and along Mayor it is abundantly indicated topographically (pl. 19c). Sierra Mayor and Cerro Borrego meet Bryan's definition of a typical sierra as being a mountain which is "longer than it is broad and presents to the observer, when viewed from the side, a jagged crest, which decreases in altitude from a point near the middle toward each end"85 (pl. 20c). They further qualify by possessing slopes scored by narrow gorges or canyons, between which are projecting ridges. The residual surfaces of the granitic masses are scored in the characteristic inverted "V" pattern, with systematically distributed cliffy slopes indicative of faulting. In this they stand in striking contrast to the Cucopa Hills, where the heterogeneity of rock and structure has resulted in the development of a very irregular pattern. Along the base of the whole range are detrital sands and gravels which display gently convex depositional profiles about the mouths of the larger reentrants, and then merge into the general level of the piedmont apron. Along the eastern front of the range the apron terminates abruptly in a scarp face, and unfailingly shows dissection. Along the western side of the range are a number of well developed alluvial fans whose slopes merge naturally with the floor of Laguna Salada. But even these are densely covered with mesquite and give the impression of being quiescent, or only occasionally active. The Cucopas are not of sufficient height to induce an appreciably greater precipitation; therefore, they share the general aridity of the desert within which they lie. Their altitude is sufficient, however, to bring something of a temperature reduction, as is indicated by an occasional snow cover on the higher peaks. This certainly means a reduced evaporation rate, with its possible significance for plant life. But few springs are known in the area. There are several along the western base of the range, one of which is hot. High up on Borrego Peak there is a small valley which contains a permanent spring, the 85 Kirk Bryan, Erosion and sedimentation in the Papago country, Arizona, Contributions to the geography of the United States 1922, U. S. Geol. Surv., Bull. 730:19-90, 1923:90.
218 University of California Publications in Geography [VOL. 5 water of which supports a grove of native palms ( Neowashingtonia). This is the only known occurrence of palms in the range. A portion of the precipitation is retained by rock tanks or tinajas. The most significant ones are those formed by irregularities in the beds of the ephemeral streams. The larger ones are found at the foot of falls where the dropping water erodes a basin in the solid rock. Others are formed through the erosion of a section of bedrock less resistant to abrasion than the rocks which surround it. Still others are found in coarsely jointed rack where the joint plane dips upstream. Through the removal of a portion of the rock between two joint planes there is formed a rock tank. Very few of these tanks retain water throughout the year. Several days spent about Sierra Mayor in midsummer failed to reveal one that contained water. The fact that wild sheep were seen on the mountain may mean that the tanks containing water simply were not encountered. The vegetation cover of the Cucopas is of a character most nearly related to that of the arid areas of Arizona and Sonora. Owing to its isolated position and to its geologically recent island or peninsular form, the area contains a number of endemic plant species
.86 In general it is characterized by the creosote bush-ocotillo-cactaceae association common to the rocky areas of the whole delta region. On the rocky aprons at the foot of the slopes are found the members of this association, together with the palo fierro, and occasionally mesquite in the sandy washes. Terebinthus, Gaertneria, the prickly pears (Opuntia bigelovi and 0. prolifera), Cactus, and Echinocactus, according to MacDougal, find suitable habitats to within a few feet of the summit at 3500 feet.87 Near the summit elevations there appear a bunch grass and a maguey, possibly Agave consociata. Just why this maguey is so restricted in distribution is not clear. It is highly valued by the natives, both as being edible and as furnishing the base for a distilled liquor. Its absence from the lower slopes may represent a removal by man of the individuals growing there. It does not, however, appear on the Peninsular Range until elevations of around 2500 feet have been reached; possibly its habitat is restricted by climatic conditions. The small but distinct difference in temperature existing between the base and summit of the Cucopas might, therefore, be sufficient as a determining factor in its range. 86 D. T. MAEDOUGAL, Botanical features of North American deserts (Washington, Carnegie Inst., 1908) :41. 87 Ibid., 41.
1932] Kniff en: The Natural Landscape of the Colorado Delta 219 THE SIERRA DE LAS PINTAS This natural area includes the irregular mass of volcanic forms bounding the Gulf Plain on the west, and is really marginal to the delta region. A few of the volcanic outliers are surrounded by the silt of the Gulf Plain. Others are isolated in the great plain of arkosic sands and gravels derived from the granitic ranges to the west. With the exception of the more ephemeral delta forms these late Tertiary extrusives are the youngest features of the landscape. It is possible that the area is still active volcanically. On the night of June 9, 1928, there was visible from El Doctor, on the Santa Clara Mesa, a phenomenon hard to explain on any basis other than that of volcanic eruption. In the sky, slightly to the west of south, there appeared a red glow which became successively weaker and stronger. Through night glasses there was visible a pall of smoke, and on one occasion there appeared what resembled a lava stream breaking over a crater rim. A rough bearing placed the site in line with the volcanic area west of the gulf. The horizon in that direction appeared to be somewhat smoky on the following day. The general region, the strong improbability of human agency, the appearance of the phenomenon itself, all favor an explanation based on volcanic activity. Without doubt, the great mass of these extrusives is much older, but they are sufficiently young that denudation and erosion have not yet destroyed the individual character of the component parts. The surface of this whole area is marked by a weird assembly of diverse forms: cones, stocks, lava-capped mesas, and flat-lying and tilted flows (pl. 25a, 29c). About the base of these mountains there are great accumulations of detritus whose profiles are adjusted to various angles of repose, suited to the nature of the constituent materials. As a whole this material is poorly rounded, though it is occasionally cut by steepwalled erosion channels, and to some extent has been carried out onto the plain to lie in attentuated rows, beveled by an ancient sea. Water resources are meager. There are reputed to be a few springs of poor water ; even the tinajas are few. The watersheds are small in area and extremely irregular in form. About some of the washes which lead out from the larger mountains there are quite convincing appearances of the proximity of water. The growth of palo fierro,
220 University of California Publications in Geograph,y [Vol.. 5 palo verde, and mesquite is heavy; bird and mammal life is well represented. Yet an ascent of several of these washes in midsummer failed to reveal a single spring or tank, and attempts at digging were fruitless. Possibly the animals get their water from a distance; the flora must depend upon the very occasional rains. The aridity and lack of soil of the residual surfaces is reflected in the scanty vegetation cover. About the lower, detrital slopes there may be a few shrubs; the higher slopes are frequently barren.
EXPLANATION OF PLATES
UNIV. CALIF. PUBL. GEOG. VOL. 5
1 KNIFFENJ PLATE 19
a. Two of the group of active mud volcanoes in the vicinity of Cerro Prieto. In the distance, to the west, can be seen the outline of the Sierra de los Cucopas.
b. A weathered remnant of an extinct mud volcano, along the edge of Sta.
Clara Mesa not far from Mesa Andrade. It will be noted how the material
tends to preserve a steep slope. Such forms as this, together with hot springs
and sulphur springs, mark the extension of the San Jacinto fault, south along
the mesa edge.
c. The steep east-facing escarpment of a section of Sierra Mayor. In the foreground appear patches of schist which probably represent the remnants of a mass intruded by the Mayor granite at the time of its formation as a laccolith.
UNIV. CALIF. PUBL. GEOG. VOL. 5
I KNIFFENI PLATE 20
a. A lagoon about five miles south and west of San Luis, Sonora. It covers
several acres and has a maximum depth of perhaps six feet. It represents an
old course of the meandering Colorado, but at present it is connected with the
main river only by high water channels.
b. A section through the river terrace near Pilot Knob. It is irregularly stratified, heterogeneous in composition, and shows such features of structure
as cross-bedding and lenticular inclusions. Without a distinguishable break
this terrace merges with the so-called mesa which marks the "ancient beach line" surrounding the Salton Sink.
c. A development of mud cracks in a small depression which permits settling
out of the finest constituents of the river alluvium. At the time this picture was taken the ground was still damp and the surface "rubbery."
UNIV. CALIF. PUBL. GEOG. VOL. 5
KNIFFEN] PLATE 21
a. Shallow mud cracks which indicate that the soil in which they are formed contains some of the coarser alluvial constituents. The sections are readily
crushed into dust and may then be removed by the wind.
b. A funnel-shaped depression in a surface recently flooded. It probably
indicates an escape of the impounded waters to a more permeable stratum below.
c. A dense growth of arrowweed in a solid stand, on land subject to annual
inundation. There are many stands much heavier than this.
UNIV. CALIF. PUBL. GEOG. VOL. 5
[ KNIFFENI PLATE 22
a. Rows of willows along the west bank of the Hardy, near its junction with
the Pescadero. The trees increase in size as one ascends the bank. This site
is on the inside of a river bend; quite possibly the rows of willows mark a
progressive march toward the cut bank at the edge of the river channel.
b. Looking toward La Bolsa from the west bank of the Colorado, a few
miles south of the mouth of the Hardy. This area is well within range of the
tide. The low-lying ground marks a movement of the channel westward. It
is alternately the inside bank of the river bend and a tide flat.
o. One of the overflow channels by which water is conducted from the Hardy
into Laguna Salads. Its surface is hard and cracked, in contrast to the alkali-
encrusted soil about it. The car is standing on the road which follows the
channel. In the distance can be seen the outline of the northern extension of
the Las Pintas range.
UNIV. CALIF. PUBL. GEOG. VOL. 5
[ KNIFFEN] PLATE 23
a. An old beach line along the eastern side of Pattie Basin, which probably
connects with the terraces on the eastern side of the Cucopas. The feature is
more evident than the picture would indicate. In the far background can be
seen the dark mass which marks a mesquite-covered fan, extending out over
the surface of the basin.
b. A section through one of the terraces extending to the eastward from
Sierra Mayor. The material is well stratified, coarse in its upper layers, and
at least fine enough below to permit it to stand in vertical cliffs. The terrace
here reaches an elevation of about forty feet above the delta surface. Not all
the terraces are of precisely the same composition.
a The terrace level, which is preserved in the distance as far as the view
is carried. This view exhibits several characteristic qualities which the ter-
races possess: a fiat top preserving a general level, an abrupt outward termi-
nation, and unfailing dissection where exposed to the occasional cloudburst
UNIV. CALIF. PUBL. GEOG. VOL. 5
I KNIFFEN] PLATE 24
PLATE 24 a. A part of the great channel excavated by New River, 1905-1907. b. Typical pattern developed in the granite masses, Sierra Mayor. The pattern exhibits regularity in its development, due to the homogeneity of the material and to the influence of jointing. a. A detail of weathering in granite as exhibited on a southeast slope of Sierra Mayor, at low elevation. The term "socket" weathering has been suggested to apply to the process by which these forms are created.
UNIV. CALIF. PUBL. GEOG. VOL. 5
I KNIFFENI PLATE 25
a. A part of the Sierra Pintas range, which illustrates the irregular sculpture
of the volcanic mountains. They are made up of a number of constituent elements, each of which tends to preserve its individual form.
b. Incipient badland topography developed in the Tertiary sediments lying
to the east of the head of the gulf. This type of topography extends inland only a few miles and then is succeeded by a high mesa, developed apparently in the same material.
o. A "desert pavement" developed in fine, well-rounded pebbles, derived largely from granite, and lying many miles from any rocky eminence. Many of these pavements are formed of much coarser material. Santa Clara Mesa.
A few creosote bushes with little sand accumulations about them interrupt the rocky surface.
UNIV. CALIF. PUBL. GEOG. VOL. 5
[ KNIFFENI PLATE 26
a. Another type of desert pavement, which derives its rock material from the slow movement down slope of the dry debris of weathering. In this case it is formed in the lee of a little basalt hill, an outlier of the Sierra Las Pintas. b. A close-up of a pavement of the same type. In this particular instance the rock was underlain by a material so fine that it has bee4 burrowed by animals. c. A north-facing barchan dune resting on a desert pavement surface, east of the head of the gulf, on the Teritary mesa.
UNIV. CALIF. PUBL. GEOG. VOL. 5
[ KNiFFEN] PLATE 27
a. Looking north from the end of the dune belt, which extends along the
western side of the head of the gulf. The sand of which the dune is composed
has come from the west, but as revealed by its profile it is subject to winds from different directions.
b. A fixed dune, showing the effects of the abrasive action of wind-transported
material. This dune is located at the southern end of the Pattie Basin, where
it is exposed to the sweep of the dominant northwest wind. To the left can
be seen the dead roots of the vegetation which held the dune together, and
which possibly contributed to its formation. The material is not pure sand,
as can be seen by the fact that it maintain a vertical face. o. Accretion dunes rising out of the barren alluvium surface lying at the
entrance to Pattie Basin. These forms afford a foothold for vegetation which
cannot grow on the alkaline soil about them.
UNIV. CALIF. PUBL. GEOG. VOL. 5
[ KNIFFEN] PLATE 28
a. Santa Clara Mesa, near the head of the gulf. A sandstone, poorly
cemented with caliche, which has been exposed by denudation. It has acted
to some extent as a more resistant layer, has been undermined, and so has
reached its present position.
b. The channels of the Colorado from the east, a short distance south of the
boundary. In the far background is the heavy forest of cottonwood trees
characteristic as a bordering strip on both banks of the old channel of the
Colorado. This was noted by nearly every visitor to leave a record of his visit.
Near this point is a very primitive ferry which serves as the sole land com-
munication, lying in Mexico, between Sonora and Baja California. o. The black-alkali surface which borders the salt beds of Laguna Salads. Under the surface the material is moist, but it dries readily on exposure to the
sun. In the distance can be seen a part of the northern range of the Cucopas.
UNIV. CALIF. PUBL. GEOG. VOL. 5
[ KNIFFEN] PLATE 29
a. The barren, alkali-encrusted surface which occupies much of the floor of
Pattie Basin. An important alteration is going on in this surface, as is revealed
by the wind-blown sand which is lodged in the little depressions. The Cucopas
in the distance, looking east.
b. Vegetation along a little sandy wash which joins with others of the same
nature to form the San Felipe Arroyo. In the middle foreground appears
garambullo (cactus), to the right creosote bush, and in the background ocotillo.
To the left can be seen a drooping mass of mistletoe, entwined about the dead
branches of the tree to which it clings.
o. The ba-salt-capped mesa which constitutes the northern portion of the
Sierra de las Tinajas.
UNIV. CALIF. PUBL. GEOG. VOL. 5
[ KNIFFEN] PLATE 30
a. At the southern end of Pattie Basin. Little granitic hills, nearly sub-
merged in a plain composed of alluvial sand and gravel, and beveled by a Pleistocene sea.
b. Extensive salt-grass plain near the mouth of the Colorado, subject to
annual inundation. Grazing cattle can be distinguished in the distance.
o. Part of the mesquite grove which marks the site of Pozo Cenizo. The
principal well is to be found in the right foreground, near the small mesquite.