Field-scale watershed evaluations on deep-loess soils: II. Hydrologic responses to different agricultural land management systems.
Key words: Baseflow, conservation practices, hydrology, runoff sediment, watershed.
The Iowa and Missouri Deep-Loess Hills is a region that is particularly sensitive to erosion and modifications in hydrology when exposed to row-crop agriculture. The region, also known as Major Land Resource Area 107 (USDA 1981), covers an area of more than 53,000 [km.sup.2] (20,500 [m.sup.2]) in western Iowa and northwestern Missouri. The western margin of the region is sharply defined by steep bluffs adjacent to the Missouri River flood plain. The eastern margin is diffusely defined by a transition to thinner loess to the east. Soils in the region are mainly Udolls and Orthents that are generally deep and range from medium to moderately fine-textured. Hapludolls, like the Monona series in the study area, also formed on the well-drained and rolling landscapes (USDA 1981). The region in Iowa is a major landform region classified by Prior (1991) as the Loess Hills.
The headwaters of streams in the region are typified by the terrain and hydrology found in the study area near Treynor, Ia. The unique topography of the region is partly a result of wind deposition of dominantly silt-size particles derived from the ancient Missouri River valley during Pleistocene glaciation, and partly a result of subsequent water erosion of these materials.
Two distinct periods of loess deposition are represented in deposits of the region. The Peoria Loess of Late Wisconsinan age (Kay and Graham 1943) is the surficial stratigraphic unit (Figure 1) in the uplands of western Iowa. Holocene alluvial deposits of the DeForest formation are the surficial unit in valley positions. The Peoria Loess thins to the east, away from the Missouri River source (Bettis 1990). It includes several paleosols, separated from the underlying Pisgah formation in which the prominent Farmdale soil was formed (Bettis 1990). Both the Peoria and Pisgah were deposited in the Late Wisconsinan, about 31,000 to 12,500 yr before present (Prior 1991).
The difference between the surface features formed on the loess deposits in western Iowa and those of similar age on other continents, is that episodic water erosion shaped the terrain as it appeared before tillage for crop production (Bettis 1990). Consequently, materials eroded from slopes and uplands were locally redeposited, forming slopewash, colluvial, and alluvial deposits on the toeslopes and in the valleys of even first-order-stream watersheds, such as those studied at this site. An older unit, the Loveland Loess, is the most widely recognized pre-Wisconsinan loess in the Midwest (Bettis 1990). The estimated time of deposition of the Loveland Loess in the area of the study is about 89,000 yr before present (Norton and Bradford 1985).
Beneath the loess deposits in much of the study area is the Yarmouth soil formed on the Pre-Illinoian till. The Yarmouth soil in this area formed in an enigmatic clay, informally referred to as the Yarmouth clay. This dominantly clay paleosol retards downward water movement. Recent coring for this study shows that in the valley positions of the watersheds, the Yarmouth has been removed by erosion, which puts the underlying PreIllinoian till in contact with alluvial and colluvial deposits.
The column of loess, colluvium, and alluvium provides a medium for relatively continuous storage and flux of water that supports perennial baseflow to streams. The hydraulic conductivity of the complex of loess, paleosols, and colluvial and alluvial deposits derived from loess are much greater [10.sup.-5]m [s.sup.-1] (3.3 x [10.sup.-5] ft [s.sup.-1])] than that of the underlying Yarmouth clay [([less than] [10.sup.-11] m [s.sup.-1] (3.3 x [10.sup.-11] ft [s.sup.-1] and Pre-Illinoian Till [([10.sup.-6] m [s.sup.-1] (3.3 x [10.sup.-6] ft [s.sup.-1])]. The only exception is in valley positions where the Yarmouth clay is missing and the till is highly fractured and weathered. This difference in hydraulic conductivity creates a zone of saturation in the loess above the Yarmouth that provides baseflow to the streams in each of the watersheds. Based on geophysical surveys (Sendlein et al. 1970), the groundwater divide for this base flow is assumed to be identical to the surface water divide as mapped in each watersh ed (Figures 2 and 3). Terrace berms, diversions, road cuts, and field lanes influence the surface water drainage area, providing potential changes in the balance between runoff and infiltration that ultimately affects base flow.
Increased tillage for row-crop production as a replacement for grazing and cattle production has increased concern regarding soil erosion by water and its effects on the balance between runoff and infiltration in this region. Management of the terrain, tillage, and crop systems in this region were hypothesized to have a measurable effect on the balance of runoff and baseflow, as well as sediment discharge. Therefore, a long-term program to study the hydrologic responses to various land and crop management systems was initiated by the Agricultural Research Service in 1963 (Karlen et al. 1999). Consequently, more than 30 yr of precipitation, runoff, baseflow, and sediment discharge data are available for four field-size watersheds representative of upland areas in the Loess Hills region where the hazard of erosion is greatest.
The objective of this report is to characterize runoff, baseflow, and sediment discharge for three landscape management systems on four small watersheds in the Loess Hills. Knowledge of significant differences in the hydrologic responses to these management practices should be useful in planning conservation and agricultural management systems to reduce soil loss and improve water resources management in this potentially fragile region.
Methods of data collection and analysis
Each of four field-scale watersheds was instrumented in 1965 to measure precipitation, runoff, baseflow, and sediment discharge. The watersheds are located near the town of Treynor in southern Pottawattamie County in southwest Iowa. They are situated as two pairs of watersheds (Figures 2 and 3) that have some common divides. Watersheds 1 and 2 are located about 5 km (3 mi) south of Watersheds 3 and 4, but all four provide drainage to first-order streams typical of headwaters in the Loess Hills. This rolling topography is characterized by gently sloping ridges, steep sideslopes, and well-defined alluvial valleys with incised channels that terminate at an actively-eroding headcut. Slopes are commonly 2 to 4% on the ridges and valleys, and 12 to 14% on sideslopes. The well-drained soils have textures of silt loam and silty-clay loam and are classified as highly erodible land (HEL). Sheet and rill erosion are serious problems with tilled and cultivated land on these soils that require adequate conservation pract ices. Grassed waterways in all watersheds have reduced the areas affected by intense gully erosion, although some gully erosion occurs on the slopes and at the headcut as it migrates up-gradient.
Major changes were made to land management systems in Watersheds 3 and 4 during 1972. Therefore, to avoid confounding hydrologic responses during the watershed management transition from 1965 to 1971, the analyses in this report only include measurements from 1974 through 1995. The term "runoff" is defined as that part of the total water discharged from a watershed in direct response to rainfall or snowmelt. Baseflow is the stream discharge between runoff events that is maintained by groundwater seepage. Sediment discharge is the total sediment discharged from a watershed. It includes the combined effects of sheet, rill, and gully erosion, and toeslope and grass waterway deposition.
Stream discharge was measured continuously with broad-crested V-notch weirs and stage recorders positioned in the gully channels. Thus, stream discharge includes baseflow from groundwater seepage into the incised channels and runoff from storm events. Baseflow volumes were determined from tabulations and semi-logarithmic plots of daily discharge. During times of runoff, baseflow was estimated from trends measured before the initiation of runoff and after recession of runoff. Runoff was calculated by subtracting the estimated baseflow from the total discharge. Runoff and baseflow were divided by the watershed area to provide unit-area values (mm) for comparison among watersheds. Precipitation was measured by recording rain gauges located on the perimeter of each watershed.
Sediment concentration samples were collected during runoff events at each weir using manual and automatic samplers. Samples were taken at intervals of 3 to 10 min during runoff events, with the larger interval used during the smaller discharges. Sediment discharge was computed from concentration and total stream discharge rates, then summarized by runoff event and day. Sediment discharge was divided by the watershed area to compute unit-area values (Mg [ha.sup.-1]).
Changes were made in the management systems used in Watersheds 3 and 4 during 1972 in an attempt to study the differences in erosion and hydrology resulting from different land management systems. Prior to 1972, Watersheds 1, 2, and 4 were planted in continuous corn with a conventional system of tillage and cultivation, and Watershed 3 was in bromegrass with controlled grazing. An extensive discussion of the management systems used in the watersheds is included in Karlen et al. (1999) and a summary is shown in Table 1.
Watersheds 1 and 2 were managed similarly and provide controls against which the modifications in Watersheds 3 and 4 can be compared. Watersheds 1 and 2 were continuously cropped to corn with conventional tillage on the contour. No conservation structures or conservation tillage systems have been used on Watersheds 1 and 2, except for grass waterways from 1965 through 1995. Tillage included moldboard plowing or heavy disk harrowing in mid-April, followed by disk and harrowing prior to planting about 2 wk later. After planting with a four-row planter, one or two cultivations for weed control were performed during the growing season. All these field operations were performed along the general contour of the field slopes.
Watershed 3 was converted to a ridge-till plant system of continuous corn in 1972. The ridge-till system consisted of early May planting along the general field contour on residue-cleaned ridges from the previous crop year. One or two cultivations were performed with a four-row cultivator to both control weeds and reconstruct ridges along the corn rows. Two cultivations were performed in most years when weather permitted, producing ridges that were about 10 to 15 cm (4 to 6 in) high.
The slopes of Watershed 4 were modified by constructing parallel terraces (Figure 3) with surface inlets and drains. The terraces were separated by about 89 m (290 ft), or twice the 1972 standard design used by the Natural Resources Conservation Service for row-cropped fields with 14% slopes (Spomer et al 1981). The surface water drainage system was designed to drain up to 5 cm (2 in) of runoff from the contributing areas in one day. The tillage management system was ridge-till, like that used in the non-terraced Watershed 3, except that the field operations followed the parallel terrace pattern rather than conforming to the general field contours.
Hydrologic responses were examined for the time periods between crop stages as well as for annual totals and maximum runoff events. These time periods are defined by specific cultural operations and crop-growth stages each year. Six periods are defined to represent uniform ground cover conditions and management effects as described by Wischmeier and Smith (1978). These periods are:
F--(rough fallow) from primary spring tillage to secondary tillage and planting;
SB-(seed bed) from planting to 10% canopy cover;
P1--(establishment) from the end of SB until crop has developed a 50% canopy cover;
P2--(development) from the end of P1 until 75% canopy cover is reached;
P3--(maturing crop) from end of P2 until crop harvest; and
P4--(residue or stubble) from harvest to primary spring tillage.
P1 and P2 were combined and redefined as a 30-day period after first cultivation. Dates for primary tillage were defined for Watersheds 3 and 4 based on cultural operations in Watersheds 1 and 2 because primary tillage was not used in Watersheds 3 and 4. The combination of seasons F (rough fallow) through P2 (crop development) were selected for comparisons among watersheds because this is when the soil is most exposed and vulnerable to erosion from precipitation. During this period, which includes about 24% of the year, an average of 40% of the annual precipitation (800 mm) occurred. This precipitation, combined with exposed soil, has the potential to produce the largest erosion and sediment delivery from the watersheds. During the F through P2 periods, an average of 88% of the annual sediment discharge occurred.
An exploratory analysis was conducted using all data for response variables. Most variables were skewed from the normal distribution. Consequently, variables were transformed by natural logarithm (lognormal) or by square root (Poisson) for an analysis of variance (ANOVA).
The ANOVA model used was a one-way fixed-effect structure in which each watershed was considered as a qualitative treatment (Mead 1988). Each year's data were considered independent samples, not replicates. Second- to fourth-order polynomials in time were added to the ANOVA model as covariates. With the covariates added, serial correlation was not a serious problem. Single degree-of-freedom contrasts were used to test differences between individual pairs of watersheds. All comparison results are reported with the probability level to quantify the significance of differences between variables measured in the watersheds.
Results and discussion
Watersheds 1 and 2 were both included in the analysis to show some measure of the variability among watersheds with the same management systems. The hydrologic responses in these two watersheds were not significantly different based on analysis of the annual total runoff (P [less than] 0.66, Figure 4) during the years 1974 to 1995. This means that there is a 66% likelihood that differences were due to chance. Similarly, no significant differences were found in annual sediment discharge (P [less than] 0.53), and little significant difference was found in annual baseflow (P [less than] 0.12) during this period. The median annual runoff was 54 mm (2.1 in) for Watershed 1, and 48 mm (1.9 in) for Watershed 2. The median annual sediment discharge was 10.88 Mg [ha.sup.-1] (4.85 t [ac.sup.-1]) for Watershed 1, and 12.34 Mg [ha.sup.-1] (5.51 t [ac.sup.-1]) for Watershed 2. The median annual baseflow was 104 mm (4.1 in) for Watershed 1, and 124 mm (4.9 in) for Watershed 2. Median values are reported to illustrate the magnitude of the difference in central tendency of the variable. The reported probability values for single-degree-of-freedom contrasts, however, were based on linear combinations (a simple difference for a given pair) of the transformed mean values.
Given the similarities in hydrologic responses in Watersheds 1 and 2, the following discussion will involve comparison of only Watershed 2 to Watersheds 3 and 4, although comparisons to Watershed 1 produce similar results. Watershed 2 was selected because it is closer in area and has a shape more similar to 3 and 4 than Watershed 1, which is also more elongated than the other three (Figures 2 and 3). However, illustrative plots include Watershed 1 as well, to display the differences between 1 and 2 for each variable.
Runoff. Watersheds with conventional rillage (1 and 2) had larger median annual runoff than either of the watersheds that were under conservation practices (Figure 4). Differences in annual runoff among the watersheds were statistically significant. Annual runoff for Watersheds 3 and 2 were significantly different (P [less than] 0.01), but the difference between Watersheds 4 and 2 was of smaller significance (P [less than] 0.26). Annual runoff for Watersheds 3 and 4 were significantly different (P [less than] 0.14). These results indicate that annual runoff was significantly less from the ridge-till watershed (3) than from either the conventional tillage (2) or the terraced watershed (4). Also, annual runoff as a percent of rainfall (Figure 5) was significantly less from the conservation watersheds [(3, P [less than] 0.01) (4, P [less than] 0.17)] than for conventional tillage (2).
Runoff is particularly important during the season between rough fallow and crop development stages. This is when the soil surface is most vulnerable to erosion resulting from exposure due to tillage and less than 75% crop canopy cover. Comparing values of this variable, the ridgetill watershed (3) had significantly less seasonal runoff than the terraced Watershed 4 (P [less than] 0.2, Figure 6). Seasonal runoff from Watersheds 3 and 4 were both significantly less than the conventionally-tilled Watershed 2 [(3, P [less than] 0.032) (4, P [less than] 0.34)]. It is clear that the difference between the terraced watershed and the conventionally-tilled watershed is less significant than the difference between the ridge-till and conventional-tilled watershed.
There also was a significant reduction in the annual peak discharge for ridge-tilled and terraced watersheds (P [less than] 0.01 for both Watersheds 3 and 4) compared to conventional tillage (2), as shown in Figure 7. This is a hydrologic response related to reduced erosion. There also were significantly smaller annual peak discharges from the terraced watershed (4) than from the ridge-tilled watershed without terraces (3, P [less than] 0.04).
Sediment discharge. Substantially smaller annual sediment discharges were recorded in watersheds with conservation tillage systems (3 and 4, Figure 8). The median annual sediment discharge was 0.96 Mg [ha.sup.-1] (0.43 t [ac.sup.-1] in the ridge-tilled watershed (3); 0.69 Mg [ha.sup.-1] (0.31 t [ac.sup.-1] in the terraced watershed (4); and 12.33 Mg [ha.sup.-1] (5.51 t [ac.sup.-1]) in the conventionally-tilled watershed (2). The annual sediment discharge was not significantly different between the two conservation watersheds (3 and 4, P [less than] 0.7), but both watersheds yielded significantly less sediment (P [less than] 0.01) than the conventionally managed watershed (2).
Substantial reductions in sediment discharged during the largest annual runoff event also were seen for watersheds with conservation practices (Figure 9). The terraced watershed (4) produced a smaller amount of sediment when compared to the conventionally-tilled watershed (2, P [less than] 0.01), but the ridge-tilled watershed (3) also produced significantly less sediment (P [less than] 0.01) when compared to the conventionally-tilled watershed (2). The differences between Watersheds 3 and 4 were not significant. Substantial differences also were measured in the sediment discharged during the rough-fallow to crop-development stage for watersheds with conservation tillage, compared to those with standard tillage (Figure 10). These reductions were significant for both conservation watersheds (P [less than] 0.01 for 3 and 4).
Baseflow. Differences in baseflow responses among the watersheds were the inverse of differences seen in runoff (Figure 11). Baseflow was significantly larger under both the ridge-till system (3, P [less than] 0.01) and the terraced ridge-till system (4, P [less than] 0.03) than under conventional tillage (2). Baseflow from Watershed 3 was larger than Watershed 4, although nor significantly larger (P [less than] 0.5). This suggests that there may be conditions under which more infiltration may occur without terraces.
A similar pattern of differences is seen when baseflow was computed as a percent of precipitation (Figure 12). This reinforces the observation that conservation tillage (3 and 4) contributes to greater infiltration than conventional tillage practices represented by Watershed 2. The baseflow from the terraced watershed (4) was less than that from the non-terraced watershed (3, P [less than] 0.33). However, this difference is not strongly significant and may be attributed to runoff from tiles draining the terraced watershed (4) rather than infiltrating and ultimately contributing to baseflow. The pattern of baseflow among the watersheds during the season between the rough-fallow and crop-development stage (Figure 13) shows significantly larger baseflow under ridge-till (3, P [less than] 0.01), and ridge-till and terraces (4, P [less than] 0.05) when compared to conventional practices (2). No significant difference in seasonal baseflow (rough-fallow to crop-development stages) was found between the ridge-till (3) and terraced (4) watersheds.
Analysis of long-term measures of runoff, baseflow, and sediment discharge showed that conservation management systems improved important hydrologic responses in the Loess Hills of western Iowa. These conservation management systems produced smaller amounts of storm-driven runoff and significantly more baseflow resulting from increased infiltration and subsequent groundwater seepage.
Accompanying this change in the balance between runoff and infiltration was a significant reduction in sediment discharged annually, as well as during peak storm events. Sediment discharge also was reduced during the season between spring tillage and the development of 75% canopy cover. It is interesting that the watershed with ridge-till and terraces did not show as much reduction in annual runoff, seasonal runoff, and seasonal sediment discharge as did the watershed with ridge-till and no terraces. These terraces were separated by 69 m (230 ft), or twice the standard distance for those slopes, and had surface inlets and drains. However, the lack of significant improvements for this terrace system suggests that ridge-till systems can produce equivalent and possibly greater improvements in hydrologic responses over conventional tillage and some terrace designs. While the significance of this relation is not large, it has implications for the efficacy of terraces in this region.
From these conclusions it is evident that land-management practices, such as ridge-till and terracing, provide multiple benefits to soil and water management. The reduced magnitude of runoff resulting from maximum storm events decreases the loss of sediment derived from sheet and ril erosion. The reduction in annual runoff and accompanying increase in baseflow provides for additional potential moisture available for crops. This is particularly important in these relatively well-drained soils.
L.A. Kramer is with the Deep Loess Research Station, Council Bluffs, Ia.; M.R. Burkart, D.W. Meek, R.J. Jaquis, and D.E. James are with the National Soil Tilth Laboratory, Ames, Ia. Both agencies are part of the U.S. Department of Agriculture Agricultural Research Service (USDA-ARS).
Bettis, E.A. III. 1990. Holocene Alluvial Stratigraphy of Western Iowa, Midwest Friends of the Pleistocene 37th Field Conference Guidebook, May 12-13, 1990. Iowa Department of Natural Resources, Geological Survey Bureau. Guidebook Series Number 9.
Karlen, D.L., L.A. Kramer, D.E. James, D.D. Bishler, TB. Moorman, and M.R. Burkart. 1999. Field-scale watershed evaluations on deep-loess soils: 1. Topography and agronomic practices. Journal of Soil and Water Conservation 54:693-704.
Kay, G.F., and J.B. Graham. 1943. The pre-Illinoian and post-Illinoian Pleistocene geology of Iowa. Iowa Geological Survey. Annual Report 38:1-262.
Mead, R. 1988. The design of experiments: Statistical principles for practical application. Cambridge University: United Kingdom.
Norton, D.L., and J.M. Bradford, 1985. Thermoluminescence dating of loess from western Iowa: Soil Science Society of America Journal 49:708-712.
Prior, Jean C. 1991. Landforms of Iowa. University of Iowa Press: Iowa City, IA.
Sendlein, L.V.A., W.P. Staub, and D.C. Kent. 1970. Geologic investigation of the USDA-ARS-SWC, Treynor, Iowa Research Watersheds, Final Report, Contract USDA 12-14-140-1906-94.
Spomer, R.G., A.T. Hjelmfelt, Jr., and R.F. Piest. 1981. Conservation aspects of selected tillage systems on western Iowa cornfields. In: American Society of Agricultural Engineers. Proceedings. of Conference on Crop Production with Conservation in the 80s September 1-2, 1980. American Society of Agricultural Engineers (ASAE) Publication 7-81: 216-227.
USDA 1981. Land Resource Regions and Major Land Resource Areas of the United States. Agriculture Handbook 296. U.S. Department of Agriculture, Soil Conservation Service: Washington, DC.
Wischmeier, W.H., and D.D. Smith. 1978. Predicting Rainfall Erosion Losses: A Guide to Conservation Planning. Agricultural Handbook 537. U.S. Department of Agriculture: Washington, DC.
Watershed tillage and conservation practices 1972-1995. Watershed 1 Watershed 2 Watershed 3 Crop corn corn corn Tillage conventional (plow, conventional (plow, ridge till and disk and harrow, disk and harrow, cultivation and cultivation) and cultivation) Conservation contour till and contour till and contour till and practices plant, plant, plant grass waterways grass waterways grass waterways Area in ha 27.6 (68.2 ac) 34.4 (85 ac) 43.3 (107 ac) Watershed 4 Crop corn Tillage ridge till and cultivation Conservation parallel terraces with practices surface inlet drainage, till and plant, grass waterways Area in ha 60.7 (150 ac)
|Printer friendly Cite/link Email Feedback|
|Author:||Kramer, L.A.; Burkart, M.R.; Meek, D.W.; Jaquis, R.J.; James, D.E.|
|Publication:||Journal of Soil and Water Conservation|
|Date:||Sep 22, 1999|
|Previous Article:||Field-scale watershed evaluations on deep-loess soils: I. Topography and agronomic practices.|
|Next Article:||Field-scale watershed evaluations on deep-loess soils: III. Rainfall and fertilizer N use efficiencies.|