L? s 5-1 1 1o lation of Water gh Pullman Soils Texas High Plains Texas A University The Texas gricultural Experiment Station H. O. Kunkel, Acting Director, College Station, Texas Percolation of Water Through Pullman Soils Texas High Plains V. S. Aronovici* Summary Nine locations including both irrigated and nonirrigated lands were explored for moisture penetration beneath the Pullman clay 10am. It was concluded that, except for those irrigated areas with an extended intake opportunity time, little or no measurable deep percolation occurs. The Pleistocene sediments transmit water readily when saturated or near field capacity, but the transmissibility is negligible with lower moisture contents. Continued productivity of the irrigated soil profile without the buildup of salinity is primarily due to the excellent quality of the Ogallala water with a favorable calcium: sodium ratio and low total dissolved salts. It should be emphasized that this study was confined to the hardland soils, and findings may not apply to areas where the soil profile is more permeable. The information given herein is for research purposes only. Reference t0 commercial products or trade names is made with the understanding that no discrimination is intended and no endorsement by The Texas Agricultural Experiment Station or the U. S. Department of Agriculture is implied. 2 VALUATION OF‘ DEEP PERCOLATION and irrigation Water through the meable subsoil of the Pullman clay ‘ related soils is important to water a; planning on the Southern Great Plai soils occupy more than 5 million ac e Texas High Plains of which nearly 3 m' are irrigated with water derived from: lala aquifer. Approximately 90 perc irrigated land is cropped to sorghum, some cotton with an average annual cation of about 18 inches. Deep perco, influence the accumulation of salts ai nitrate with implied aquifer pollutiof might provide a source of recharge to . diminishing ground water reserve. ’ Theis (9) estimated that annu, to the Ogallala is less than 0.5 inch. H however, that recharge occurs only in a ’ the Ogallala is either exposed or y porous sediments. Cronin and Myers} out that in the areas of slowly pe I underlain by relatively impermeable cal'_ little, if any, deep percolation will a f gated land as a contributor to deep l is not mentioned. ’ The evapotranspiration potential f f far exceeds the average annual or mon" pitation. Taylor et al. (8) show that, oration exceeds precipitation in every-- three times or more and on an annu, approximately five times. Occasiona exceeds evapotranspiration when sto over a period of several days or are, by short intervals. Evapotranspiraf stantially reduced by cloud cover bet mittent storms. Antecedent irrigatio fall may fill the soil water reservoir J addition of water from a storm or irri result in runoff and possible deep peri Taylor et al. (8) measured infilt ; into the Pullman clay loam. An initv, 0.62 inch per hour was obtained foll‘ filling of a level border but rapidly d the soil reservoir filled, to a final p inch per hour. In a 15-hour period, of water entered the profile. Thus, prolonged irrigation applications or J and continuing rainfall will the soil wa capacity be exceeded and deep percolati occur at a rate of 0.05 inch per hour. p, Since only limited measurements o ture below the 6-foot depth have been area, deep percolation that may have w; not been investigated adequately. of this study was to determine the a deep percolation occurring under sevc _ practices in irrigated and nonirrigated j *Former research soil scientist, USDA ‘i Great Plains Research Center at Bushland, f Research Service, U. S. Department of Agric land, Texas. PROCEDURE of Study Area y sites were located on or immediately to the USDA Southwestern Great Plains v Center at Bushland. All sites were 0n an clay loam profile described by Taylor ,. This soil is underlain by a thin porous yer which grades into buff to reddish i-ipluvial sediments of Pleistocene age. iments vary in thickness, depending j elevation, from less than 25 to more i feet. A sometimes-indurated caprock r separates the Pleistocene from the un- dewatered upper strata of the Ogallala ‘n.- The Pleistocene sediments form a e blanket over the Ogallala, and it is in rial that the present surface relief of i playa has been sculptured. Pleistocene in the High Plains area has red by most hydrologists because the ‘g is usually nonwater bearing. A limited f Work was done by Frye and Leonard F‘ e physical and hydrologic properties of iments and their mode of formation. they are the» parent material for a ‘anse 0f Pullman soils and form the *~ ntle overlying the Ogallala in this region, on added significance from the stand- aquifer recharge of both potable and Fwater. VTSTORY OF LAND USE AT BORING SITES, PULLMAN w’ Depth Land use and , of hole Period water management ‘ 5 E1 " 23 -1917 Grassland, native 1917-1937 Wheat, dryland 1937-1969 Grassland, revegetated 51 -1917 Grassland, native 1917-1949 Wheat, dryland 1949-1955 Limited irrigation of wheat 1955-1963 Continuous basin irrigation 1963-1969 Continuous basin irrigation with no nitrogen fertilizer -1917 Grassland, native 1917-1949 Wheat, dryland 1949-1955 Limited irrigation of wheat ~ 1955-1969 Continuous basin irrigation ; 3O -1917 Grassland, native l 1917-1943 Wheat, dryland 1943-1969 Wheat-sorghum-fallow, ~ terraced l 21 1927-1958 Wheat, dryland ' 1958-1969 Wheat-sorghum-fallow, furrow irrigated 0.15% . slope -1950 Wheat, dryland 1950-1970 Wheat-sorghum-fallow, furrow irrigated 0.15% _- slope ' 44 $1950 Grassland i 1950-1968 Wheat-sorghum-fallow, furrow irrigation 1968 Sugar beets, furrow irrigation, 0.2% slope a 1969 Fallow l 49 Grassland, native - 41 Grassland, native 46 2o tube; no soil or substratum samples obtained. é} 6mg g I Yhfii/ifi‘? fiéwfiiflfi’ Figure 1. Breakdown of soil and substrata coring equipment. Seven representative sites within a radius of 1 mile were selected for this study. Table 1 sum- marizes the land use and water management of these sites. The only differences lie in the land management history and slope. Two additional moisture profiles were obtained for native grass- land using the neutron meter in access tubes established for other studies. Sampling Equipment It was necessary to design special coring equipment capable of taking 1-foot increments to depths in excess of 50 feet with minimum core disturbance or contamination. Since no water -could be used in the sampling, conventional, hy- draulic-rotary well-drilling equipment was unsuit- able. Equipment was constructed to work on a small Damco AR 500 drilling rig. The sampling unit consists of three parts—the core barrel, a thrust bearing and a reamer (Figure 1). The core barrel is 18 inches long and has a 2.7-inch inside diameter with a bit silver soldered to one end. The core barrel is attached to the thrust bearing by large set screws, which, when in place, are flush with the outside of the core barrel. The reamer is attached to the upper portion of the thrust bearing with set screws. The reamer consists of a cylinder slotted on three sides and fitted with cutting blades to ream the hole to approximately 3.2 inches in diameter. The cut- tings are collected in the reamer cylinder. An adapter fitted with standard drill pipe threads is attached to the upper end of the reamer with set screws. This adapter connects directly to the drill pipe. Samples are obtained by applying hydraulic thrust up to 6 tons as the drill pipe is rotated. As the reamer rotates above the core barrel en- larging the bore hole, the core barrel is pressed into the sediment without rotation so that an undisturbed core is obtained. Cores are taken in 1-foot increments. After the sample is taken, the assembly is pulled from the bore hole, the core barrel removed and the reamer cleaned of cuttings. A second core barrel is attached to the thrust bearing and the process repeated. The material in the core barrel is carefully removed by sliding the sample out the upper end. The few thousandths of an inch difference be- tween the inside diameter 0f the bit and the barrel allows the sample t0 slide freely out of the core barrel. After the sample is placed in a tray and the loose material is removed, the core length is measured and the core weighed. An aliquot sample is removed for gravimetric soil moisture determination and a second subsample taken for ion analyses at a later date. Moisture content in milliliters per cubic centimeter (ml /cm3) is later calculated using the dry bulk density values and gravimetric moisture determinations. Sampling continues until it is no- longer possible to press the sampler using the full 6 ton thrust. Sampling of the Pleistocene alluvium with this unit became impractical when the moisture content dropped to less than 0.18 ml/cm3. As this moisture level is at or below wilting point, there seemed to be no need to continue, as a wetting front had not extended below this point within the time of modern agricultural history. Establishment of Water Holding Characteristics of Pleistocene Without predetermination of the moisture holding characteristics of the Pleistocene, mois- ture profiles would have little significance relative to deep percolation. Data from an earlier study of ground-water recharge through the Pleistocene suggested these values (1). This study used recharge basins excavated through the Pullman soil on grassland which had never been irrigated. A set of neutron access tubes was installed in the basins before they were filled with water. These access tubes extended to the caprock 50 feet below the original ground surface. The moisture co-ntent was measured in foot increments from the surface to the caprock before water was applied to the basin. Moisture contents were recorded during recharge at or near satura- tion in a zone above the caprock and below a perched water table. Thirty-eight days after water additions were stopped, moisture was again measured. The perched Water table disappeared TABLE 2. WATER HOLDING CHARACTERISTICS AND BULK DENSITY OF PLEISTOCENE SEDIMENTS within 24 hours after the basin was em’ eating that free drainage took place. T observations suggested values for speci tion or field capacity, saturation and, - wilting point. These observations Weref supported by groups of undisturbed co ._ at the 6-foot depth near neutron access and between 6 and 12 feet at another = One-third and 15-Bar suction moisture nations were made on these samples? Results, compared in Table 2, sug the wilting point or 15-bar suction is a’ ml /cm3, field capacity or 1/3-bar suction 0.32 ml/cmi‘ and saturation will app exceed 0.40 ml /cm3. These volumetric contents may also be expressed as feet per foot of soil. With this basic info f is possible to determine, at least with'_ historical time, the depth and amount. percolation. ' The diffusivity of the undisturbed A lected near neutron access tube B-1 was using the E. J. Doering one-step techni These values were then used to calcula saturated hydraulic conductivity. A suction with a moisture content of 0. the flow was 18.9 feet per year, but. suction with a moisture content of 0. f hydraulic conductivity was reduced to A per year. The sharp drop in hydraul'- tivity when the water content decr J l/g-bar suction suggests that deep perco " dry sediments would produce a Well-de p ting front. ' RESULTS Characterization of Pleistocene Sedimentsf Field observation of the several hunr taken from the seven borings revealed t_ cene to be a porous material contai =_ root and worm hole casts ranging in p, 5 millimeters (mm) to less than 1 p ‘Unpublished data, Paul Unger, Southwestern tf Research Center at Bushland. . 2Unpu~blished data, Harold Eck, Southwestern t Research Center at Bushland. 3 Water content of sediment Calculated Source of data Saturation l/a b6" pore 72 days drainage l5 bar space ml/cma Deep percolation sampling stationsl 0.19 to 0.20 0.40 Neutron access tubes in recharge basinsz 0.40 0.33 0.19 .40 Undisturbed cores é-foot depth3 .42 .32 Undisturbed cores 6-l2-foot depth‘ .20 .42 ‘Summary of cores taken between 5 and 50 feet. 2See Aronovici et al. (l). “Unpublished data collected by Paul Unger, SWGPRC. ‘Unpublished data collected by Harold Eck, SWGPRC. 4 jre found to the maximum depth of sam- .The color ranges from almost white to -- The light color is dependent upon 'um carbonate content, and the brick-red gest fossil soil profiles or weathered sur- g nconsolidated caliche stringers are com- ith a material so highly structured, a 'onal mechanical analysis is of little signi- M The pluvial sediments are composed of e, well-rounded silica sand, probably of l origin, suspended in a matrix of silt and ‘ e material is quite stable when confined i, , but a clod placed in water disintegrates 1 ough bulk density and degree of cement- ary considerably from foot to foot, a 7;: of the average of all samples taken at tion shows the material to be quite uni- r he average bulk density of all cores taken i: g/cmi‘, and the average bulk density individual borings ranged from 1.52 to m3. Thus, comparisons of cumulative ‘isture between stations are reasonably Profiles f differences in moisture contents of the ne sediments underlying the several land treatments are best shown in a comparison of cumulative moisture values with depth. These data for all sampled profiles and the two neutron access holes drilled in grassland are summarized in Figure 2. By utilizing the data presented in Table 2, it is possible to draw a cumulative mois- ture curve for a saturated profile and a profile at 1/3-bar suction. A point at the 50-foot depth for the 15-bar suction shows a cumulative mois- ture content of 9.5 feet -of water. The B-1 cumu- lative moisture underlying grassland terminates at the same point, suggesting that the profile underlying the grassland averages 15-bar suction moisture content. Two other grassland samplings, F-1 and Boring 1, in part, have less than 15-bar moisture. It can be assumed that under present ecological conditions no deep percolation has oc- curred in grassland. The grassland moisture profiles will be used as a base for comparison. If the moisture content in the profile, particularly beneath the root zone, exceeds the 15-bar moisture content, deep percola- tion is assumed to have occurred during recent times. Where the moisture content is at or below the 15-bar value, no deep percolation is assumed. Boring 4 underlies land that has been ter- raced since 1943 and cropped to a dryland wheat- sorghum-fallow rotation. The boring was made ~—-l6 QYQ \\ l1’ e‘ ‘O ,1’ ll I’, 2-1 I,’ / -_ll0 I’ ' PROFl z’, ’/ I _ 0 AT LE ’ ,»" l5 BARS ,_/ 7?/’Z”l /, +_.g] —_8 ’/ Ill-- fi/ll/ -\~n~-‘{.F_] /b _4 Figure 2. Cumulative soil —-2 moisture in Pullman clay loam and underlying Pleis- l I I tocene sediments. 1° 20 so 4o 5o DEPTH BELOW sou SURFACE IN FEET fiNative grassland lilevel border, irrigated ‘ryland, wheat-sorghum-fallow, terraced irrigated wheat-sorghum-fallow, infrequent applications l otive grassland tive grassland rrigated wheat-sorghum-fallow, frequent applications 1.2 percent slope rrigated wheat-sorghum-fallow, frequent applications 0.2 percent slope OBORING B-1 Figure 3. Moisture distri- bution in Pullman clay loam and underlying Pleistoce-ne sediments. DEPTH lN FEET o 0 .1 .213 .4 0n the natural land slope on the normal terrace interval. This cultivated dryland has accumulated some water to 17 feet as compared with grassland. Boring 5, shown only as a point on Figure 2, is in a field that was dry farmed to wheat-fallow from 1927 to 1958 and sparingly irrigated in a wheat-sorghum-fallow sequence since 1958. Sur- prisingly, deep percolation is less than for Boring 4. Boring 6 is in a furrow-irrigated field with 1.2 percent slope. It has been frequently irrigated in a more intensive cropping system than the Boring 5 site. No evidence of deep percolation at Boring 6 was observed beyond 8 feet. In con- trast t0 Boring 6, Boring 7 is in a field with 0.2 percent slope that has been furrow irrigated and intensively cropped for the same period o-f time and by the same operator. Here deep percolation extends to more than 4O feet. Borings 2 and 3 (number 3, n-ot shown on graph, is identical to number 2) are located in a closed, level basin where irrigation and rainfall since 1955 have provided some intake opportunity times extending from 24 to 48 hours. Here deep percolation reaches 50 feet, and the moisture content through- out the profile approaches 1/3-bar suction or field capacity. A group of representative moisture profiles is presented in Figure 3. For all practical pur- poses, 0.2 ml/cm3 is equivalent to wilting point, TABLE 3. SOLUBLE ION ANALYSES BY DEPTH INCREMENTS TO 11 BORING 2 l .1 .4 FEET, GRASSLAND SITE, BORING 1 . BORING 4 BORING 5 BORING a .. 0 .1 .2 .3 .4 ml /cm3 0.3 ml/cm3 is approximately l/g-bar suc field capacity and 0.4 ml /cm3 approaches 5 tion. Only in the case of Boring 2, which u A a heavily irrigated level-basin, does thef profile approach field capacity. B-1, und grassland, shows three distinct zones of y" above 0.2 ml /cm3. A possible explanation i these zones represent historic very wet g when downward percolation occurred. i, the downward progress of these moisture- would have been extremely slow. It is j, that these zones are associated with g exceptional rainfall seasons—probably t“ zones of finer textured material having a’; moisture content at wilting point. a Salinity Distribution A reliable indicator of downward pe ~ is the salinity concentration and distrib the profile and substrata (7). A soil sa- extract analysis of Pullman clay loam 1 never been irrigated is summarized in . Some difficulty was experienced in securi anion-cation balances; however, the analy‘ that under normal grassland conditions é a natural concentration of ions in the; between the 3- and 7-foot depth. The dist of total soluble salts as shown by the i; trical conductivity) x 103 @ 25° C of _ Cations Anions Depth EC Ca -Mg Na K Total Cl HCO3 SOi Total Inches mmhos meq/l 3- 17 .69 3.2 1.9 2.9 .4 8.4 .4 4.3 17- 24 .57 1.9 1.0 3.9 .2 5.2 1.9 3.4 .1 5.5 24- 36 .86 2.6 1.2 5.4 .2 9.4 2.0 2.4 2.8 7.2 36- 48 3.90 24.5 5.8 8.0 .4 38.8 8.8 1.2 28.7 38.7 48- 60 2.8 14.4 4.3 7.7 2.7 29.1 10.9 1.0 13.8 25.7 60- 72 3.6 24.1 5.3 7.8 .5 37.8 12.0 1.1 24.7 37.8 72- 84 2.0 8.7 2.7 7.2 .4 19.1 10.9 1.2 5.1 17.2 84- 96 1.7 8.5 2.6 7.0 .4 18.5 13.4 1.1 3.5 18.0 96-108 1.5 6.3 2.0 5.8 .2 14.8 10.9 1.0 3.9 15.8 108-120 1.4 6.8 1.9 5.8 .2 12.4 10.6 1.2 2.2 14.0 120-132 1.4 5.5 2.0 5.1 .2 11.8 11.3 1.3 5.0 17.6 6 ........... w. 50o 1 2 s 4 ECxlO3ut25°C 01234 Ln extract is presented in Figure 4. At 7;: where deep percolation is not indicated oisture profile—Borings 1, 4, 5 and 6— concentration of total salts is present in e between 3 and 7 feet. That this zone leached in Borings 2 and 7 again reflects ;. percolation has occurred. ride distribution in the soil and substrata ivide a more reliable means of tracing it movement as the chloride ion moves ,'th relatively slow moisture flow. Dyer T this technique in tracing moisture move- j. Central California. Figure 5 presents fconcentration profiles of five represen- a files. The chloride concentration has ected slightly deeper than the total salts. _ Boring 4 (dryland terrace) and Boring W irrigated on 1.2-percent slope) show ght downward deflection of the chloride tion supports the moisture profile evi- l little or no deep percolation has occur- special interest is the concentration of p‘. between 38 and 47 feet in Boring 2 . irrigated level basin). It is not clear "large concentration of chlorides is found epth while Boring 7 (furrow irrigated “rcent slope) does not show this concen- Figure 4. Electrical con- ductivity of Pullman clay loam and underlying Pleis- tocene sediments. DISCUSSION The seven borings and two neutron tube ob- servations of soil and substrata moisture distribu- tion are believed to be fairly representative of the conditions found in the Pullman clay loam and related soils of the Southern High Plains. Data from these observations support some reasonably sound conclusions. The sharp drop in diffusivity when the moisture content of the Pleistocene sediments falls below l/g-bar suction suggests that there would be a relatively sharp demarcation of a wetting front had there been deep percolation within recent time. This is shown clearly in Boring 7 at 44 feet and by the impossibility of sampling beyond 30, 21 and 20 feet at Borings 4, 5 and 6, respectively. Soluble salt profiles, as shown by the EC and chloride concentration of the saturation extracts, support the conclusion that only under very special conditions does deep percolation occur. There has been little or no deep percolation on native or revegetated grassland within historic time where natural surface drainage occurs. Plains grasses have deep rooting systems that are able to deplete soil moisture when it becomes available during the growing season. Since the soil profile underlying the grass is depleted much of the time, rainfall is insufficient to rewet the entire profile and initiate deep percolation. When ORINGA eonmo o scams 7 BORING 1 Biiilsllg 2 BTERRACE FURROW IRR FURROW ma OGRASSLAND Cinnamon ODRY LAND 6.2x SLOPE 02x SLOPE and I'- LU h“. 2o Z ; 3O Distribution of I Pullman clay z I nderlying Pleis- L3 4Q ents. 40 500 20 4O 500 2O 40 500 2O 4O Meq/l Cl‘ intense rains occur, excess rainfall runs off. The tightly matted and consolidated upper root zone may reduce the infiltration rate. The conditions are different on dry-farmed cropland. Here there are periods during fallow or the early part of the growing season when the relatively high soil moisture and a sequence of moderate storms might contribute to deep per- colation. The surface roughness of cultivated land may provide a longer surface detention period and, consequently, a longer period for downward percolation. Cultivated land also may be subject to more extensive cracking of the slowly per- meable subsoil which may contribute to deep percolation when the profile is quite dry. The moisture profile on the terraced land suggests that this may occur, but in no significant quan- tities. Deep percolation on irrigated land does not occur to any significant extent under most condi- tions as shown by Borings 5 and 6. However, under the following combinations of conditions, deep percolation can occur: 1. Frequent irrigations with prolonged ap- plication time. 2. Irrigation followed by rainfall. 3. Gentle slopes with resulting slow irriga- tion and rainfall drainage which provides for a prolonged intake opportunity time. That these conditions are conducive to deep percolation is supported by a comparison of Bor- ings 5 and 6 with Borings 2 and 7. Deep perco- lation at Boring 2 is roughly 5 feet of water for 50-foot depth while Boring 7 shows an accumu- lation of 2.2 feet of water in 43-fo-ot depth. At the time Boring 7 was sampled, soil profile moisture was very low compared with that in Boring 2. Boring 5 and Boring 6 show less than a foot percolation in 20-foot depth, and in both caseis most of the percolated water is in the upper 10 eet. The Pullman clay loam has an inherent salinity. Surprisingly, there appears to be no major buildup of salinity in the areas either spar- ingly or frequently irrigated. This condition, when compared with those in other major irri- gated areas throughout the world, is unique. There are possibly two main reasons: First, the total salt concentration of the Ogallala water is low with an EC x 103 of less than 0.5, and the calcium: sodium ratio is ideal for sustaining good soil structure; second, the total application of irrigation water is small compared to that in many other irrigated areas. Other possibilities may be considered. Crops remove some of the i-ons (6). It is estimated that sorghum may remove 10 per- cent of the salts contained in an 18-inch irrigation application. However, the primary ions removed include calcium, potassium and chlorides while sodium, magnesium, bicarbonates and sulphates are removed in very small quantities. Under these 8 circumstances it might be assumed thatl Sodium-Adsorption Ratio (SAR) W011i. ally increase. However, this was not w, both irrigated and nonirrigated fields a the same with an SAR of 3.0 or less. Under present land management a 1 tion practices, the possibility is remote y“ are any measurable quantities of deep pe_ through the Pullman qlayifloiam prQflle an lying blanket of Pleistocene sediment q Ogallala aquifer; ifdeep percolation r- will not be a significant source of rec pollution except where water is held on r for sustained periods. l‘ ACKNOWLEDGMENT i 1.‘ This study was a cooperative effo if Texas Agricultural Experiment 3133131011. Soil and Water Conservation Research Agricultural Research Service, U. S. De of Agriculture. ’ Thanks are extended to Robert Core Drill Qperator, for the design and f = < of the special coring equipment which I. study possible. ’ LITERATURE CITED 1. Aronovici, V. S., Arland D. Schneider, i» Jones. Basin recharging the Ogallala aqu f Pleistocene sediments. Proceedings, Ogall__ Symposium, Texas Tech University, Lub: pp. 182-192. May 1970. 2. Cronin, J. G., and Myers, B. N. A sum _, occurrence and development of ground Southern High Plains of Texas. United " logical Survey Water-Supply Paper 1693. 3. Doering, E. J. Soil-water diffusivity by method. Soil Science, VOL 99, NO- 5, PP 32 j 1965. “ 4. Dyer, Kenneth L. Interpretation of chloride ion distribution patterns in adjacent irriga ‘-_ irrigated Panoche soils. Proceedings, Society of America, Vol. 29, No. 2, pp 170 _, 5, Frye, John C., and A. Byron Leonard. _ Cenozoic geology along the eastern mar. High Plains, Armstrong to Howard counti of Economic Geology, University of Texas. ‘ Investigations No. 32. November 1957. 6. Morrison, F. B. Feeds and feeding. Tw The Morrison Publishing Co., Ithica, N. V. Appendix pp 963. 7. Stewart, B. A., F. G. Viets, L. Hutch ~~ Kemper, F. E. Clark, M. L. Fairbourn, and Distribution of nitrates and other pollu fields and corrals in the Middle South ' of Colorado. USDA ARS 41-134, 1967. 8. Taylor, Howard M., C. E. Van Doren, ‘Curtis _ and James R. Coover, Soils of the Southw Plains Field Station. Texas Agricultural . Station, College Station, Texas. MP-699. ‘_ 9. Theis, C. V. Amount of ground-water rec Southern High Plains. Trans. American Union, 18th Annual Meeting. pp. 564-5;