| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Oregon State Univ., Columbia Basin Agric. Res. Cent., P.O. Box 370, Pendleton, OR 97801
Corresponding author (w-payne{at}tamu.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsNOTESResultsDiscussionREFERENCES |
|---|
than by the Leggett model, which uses March soil water storage and spring rain 
. Yield prediction was improved when ET was divided by seasonal mean daily water vapor pressure deficit (
) 
. Multiple-regression equations using monthly rain predicted wheat yield well 
, but coefficients differed among data sets. A model using monthly rain and heat degree day sum (HDDS) predicted pea yield much better 
than ET-based equations 
, suggesting that pea yield is limited by factors other than water. For wheat, ET/-based models should replace the Leggett model. However, for pea, multiple regression models predict yield better than ET-based models.
Abbreviations: ANOVA, analysis of variance • DOY, day of the year • ET, crop water use (transpiration + evaporation of water from the soil surface) • Fall R-CH, fall roto-till wheat residue and chisel plow pea residue • Fall MBD-MBD, fall moldboard plow wheat residue and moldboard plow pea residue • Fall SWP-SKW, fall sweep wheat residue and skew-tread pea residue • HDDS, heat degree day sum • Spring MBD-MBD, spring moldboard plow wheat residue and moldboard plow pea residue • , mean daily water vapor pressure deficit for the growing season • WUEET, water use efficiency (crop yield/ET)
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsNOTESResultsDiscussionREFERENCES |
|---|
Simple models have been used to predict yield of wheat and pea. Winter wheat yield in this region was linearly related to the sum of soil water storage in March plus the amount of subsequent spring rain (Leggett, 1959). This model is still widely used today to predict yield potential and to estimate N requirements of crops. The sum used by the Leggett model serves essentially as a proxy for ET, which has also been linearly related to yield of wheat and many other crops (e.g., Hanks, 1983). However, the Leggett model does not take into account ET from sowing to the time of measurement of soil water storage in March or reduced crop water availability due to runoff or drainage. It also does not take into account the annual variation in atmospheric evaporative demand, which influences the relation between crop yield and ET (e.g., De Wit, 1958).
There are fewer studies on ET of pea than there are on ET of wheat. Farah et al. (1988) found pea ET to range from 350 to 500 mm yr-1. Gregory (1984) attributed lower transpiration efficiencies at Akron, CO to greater atmospheric evaporative demand. In the inland Pacific Northwest, pea has been found to be more sensitive than wheat to water availability during the growing season and especially sensitive to high temperatures near harvest time (Pumphrey et al., 1979). During May and June, even one day of excessive heat can reduce pea yield and quality (Kraft et al., 1991).
Multiple-regression models that use rain and temperature as inputs have also been used to predict yield of winter wheat and fresh pea. These models have the advantage that ET data are not required. Ramig and Pumphrey (1977) found that overwinter (Sept.–Mar.) and growing-season (Apr., May, and June) precipitation accounted for 64% of the year-to-year variation in wheat yields. Pumphrey et al. (1979) found good correlation between fresh pea yield and the rain and temperature distribution during a 40-yr period.
Fresh pea is harvested and graded based on its tenderness, as measured by a tenderometer, rather than on its moisture content or physiological maturity. The optimum tenderometer readings for pea harvest are between 90 and 110 (Pumphrey et al., 1975; Kraft et al., 1991). Near harvest, pea weight increases rapidly while tenderness decreases rapidly. This makes it more difficult for models to accurately predict yield and for agronomists to interpret experimental data because the apparent treatment differences may be due to effects upon maturity (Pumphrey et al, 1975). Pumphrey et al. (1975) introduced a method of standardizing yields to a tenderometer reading of 100, but to our knowledge, it has not been tested on an independent data set as a method of improving the precision of simple yield models or detecting treatment effects.
Soils of the inland Pacific Northwest are prone to erosion, especially during the winter, when precipitation occurs during freezing and thawing events on fields planted to winter wheat (Zuzel, 1994). During such events, wheat seedlings provide very little ground cover, infiltration rates of the frozen soils are very low, and erodibility of the intensively worked soil is high. Despite the high erosion potential in the inland Pacific Northwest, most winter wheat–fresh pea systems include intensive tillage to reduce heavy crop residues at planting and to control pathogens and weeds. A common tillage sequence for pea production includes fall plowing of wheat stubble, spring tooth harrowing, a second and possibly third harrowing, and after planting, packing the soil with a roller and attached harrow (Hoag et al., 1984). Winter wheat is also planted into intensively worked ground (Hammel, 1995; Pikul et al., 1993).
Tillage systems that maintain residue cover, especially during the winter months, are recognized as important methods of reducing soil degradation and erosion. However, it is not clear that such conservation tillage systems are as economically profitable as conventional tillage systems. For example, Kraft et al. (1991) found that a tillage program for a winter wheat–fresh pea rotation in the inland Pacific Northwest that included the moldboard plow, three cultivations, and post-sowing packing of the soil gave a larger annual return than conservation tillage systems. Similarly, Hammel (1995) found in a 4-yr study that zero tillage systems obtained only 72% of the yield of conventional moldboard plow systems, whereas conservation tillage obtained 92%.
The objectives of this study were (i) to compare the tillage effects on the yield and yield water-use relations of a wheat–pea cropping system during a 21-yr period in eastern Oregon, and (ii) to evaluate simple yield models that use monthly weather data or ET to predict the yields of winter wheat and fresh pea.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsNOTESResultsDiscussionREFERENCES |
|---|
400 mm, and the soil is a Walla Walla silt loam (coarse, silty, mixed mesic Typic Haploxeroll). From 1967 to 1991, fresh pea was grown in rotation with wheat. The experimental design was a randomized block with four replications. Each replicate contained eight plots (two crops x four tillage treatments). The location of the pea and wheat within a replicate alternated from year to year. Originally, the individual plot size was 7.7 by 36.5 m. In 1976, the east half of the experiment received 2020 kg ha-1 lime. Thereafter, only yield data from the unlimed half were used for analysis.
The wheat varieties used were Nugaines from 1967 to 1974, Hyslop from 1975 to 1978, and Stephens after 1979. From 1967 to 1981, the winter wheat received 45 kg N ha-1 as ammonium nitrate (NH4NO3) (34–0–0). This rate was increased to 67 kg ha-1 in 1982, and again to 90 kg ha-1 in 1985. Ammonium nitrate was broadcast before planting, which occurred as soon after 10 October as soil moisture was sufficient for germination and early crop growth. In 1974, winter wheat was killed by frost; therefore, wheat was planted the following spring.
Pea was planted in late March or early April and harvested in late June or early July. The variety Dark Skin Perfection was used throughout the experiment. Pea traditionally received 22.4 kg N ha-1 as either ammonium sulfate [(NH4)2SO4] (21–0–0–24S) or ammonium phosphate sulfate (16–20–0–14S).
The primary tillage treatments are summarized below. Additional details on secondary tillage operations were given by Payne et al. (2000).
Treatment 1: Roto-Till and Chisel Plow (Fall R-CH)
Wheat stubble was roto-tilled to a depth of 10 cm in August. In the spring, plots were sprayed for weeds with glyphosate [N-(phosphonomethyl)glycine], swept once with a V-shaped sweep to a depth of 5 to7 cm, and rod-weeded. Pea plots were swept to a depth of 5 to 7 cm after harvest in July. In late September or early October, they were chisel-plowed twice to a depth of 30 to 38 cm and then rod-weeded to a depth of 4 cm before seeding wheat. Residue cover in the fall following pea was 10%; residue cover in the fall following wheat was 40%.
Treatment 2: Fall Moldboard Plow and Moldboard Plow (Fall MBD-MBD)
Wheat stubble was moldboard-plowed in July to a depth of 20 cm. In the spring, plots were sprayed for weeds, tilled twice with a spring-tooth harrow to a depth of 15 cm, and roller-harrowed before seeding pea if necessary. The spring-tooth harrow was equipped with C-shaped shanks. Pea vines were moldboard-plowed in July to a depth of 20 cm, sprayed with herbicide to control weeds if necessary, tilled twice with a light disc harrow 10 cm deep, and roller-harrowed to reduce clods before seeding wheat. The residue cover in the fall following pea was 1%; the residue cover in the fall following wheat was 5%.
Treatment 3: Spring Moldboard Plow and Moldboard Plow (Spring MBD-MBD)
Wheat stubble was spring moldboard-plowed to a depth of 20 cm. Secondary tillage and management of the pea vines were the same as in Treatment 2. Residue cover in the fall following pea was 1%; residue cover in the fall following wheat was 80%.
Treatment 4: Fall Sweep and Skew-Tread (Fall SWP-SKW)
Wheat stubble was skew-treaded once or twice in March with a Durham (Durham, OH) skewtread1. The skewtread is a bidirectional, light tillage implement. As used in this experiment, it is similar to a culti-hoe and much less aggressive than a rotary hoe. In the reverse direction, it is similar to a culti-packer. Plots were swept once to a depth of 5 cm and rod-weeded before planting pea. Pea vines were skew-treaded 2 to 3 times in the summer. In the spring, plots were sprayed if necessary and rod-weeded twice. Pea vines were swept in July and again in October before wheat was planted. Residue cover in the fall following pea was 20%; residue cover in the fall following wheat was 80%.
The Spring MBD-MBD and especially the Fall SWP-SKW treatments can be considered as conservation tillage systems in terms of residue cover and erosion control, especially during the crucial winter months. Conversely, the Fall R-CH and, especially for Pendleton area, the Fall MBD-MBD treatments can be considered as conventional.
Allmaras et al. (1987) and Pikul et al. (1993) detected no differences in soil saturated hydraulic conductivity among these tillage treatments. However, bulk density profiles did differ among tillage treatments (Pikul et al., 1993). All but the Fall SWP-SKW treatment had tillage pans, but the Spring MBD-MBD treatment tended to have a tillage pan that was more compact than that of the Fall MBD-MBD treatment. Bulk density was greater near the surface for the Fall SWP-SKW treatment, but it decreased with depth. Furthermore, the surface organic matter content was greater near the surface for the Fall SWP-SKW treatment, and pH was lower (Pikul et al., 1993).
Neutron access tubes were installed in the eastern half of each experimental plot to a depth of 2.44 m or until semi-impermeable contact (calcareous hardpan or basalt bedrock) was reached. From 1967 to 1989, neutron probe measurements were taken at intervals of 31 cm in depth from three to several times during the growing season to determine soil water content. Until October 1974, a Troxler neutron probe (model 1255) and scaler rate meter (model 2651) were used (Troxler Lab., Research Triangle Park, NC). Thereafter, a CPN (model 503) neutron probe was used (Campbell Pacific Nucl., Pacheco, CA). Instrument failure early in the summer of 1979 prevented an estimation of total wheat ET.
Neutron probes were field-calibrated. For the Troxler probe, internal standard counts were taken several times during the day of measurement. Additionally, readings were taken in each of three manufacturer-provided containers with known volumetric water contents on a weekly basis. Count ratios for these manufacturer-provided containers over the 5-yr period had a coefficient of variation of <0.01. The CPN probe was field-calibrated several times during its 15 yr of use. A typical calibration had an r2 of 0.92 and a standard error of estimate of 0.0094.
Crop ET was calculated using the soil water balance equation
 | (1) |
2% slope) ground, are ignored in Eq. [1].
Pea yields are reported on a fresh weight basis, and wheat yields are reported on a dry weight basis. To calculate WUEET, yield was divided by ET. An analysis of variance (ANOVA) for pea and wheat yields was made for each year, using tillage as the only experimental factor, and again for all years, using tillage and year as experimental factors. For pea, an ANOVA was also made for yields that were corrected to 100% tenderometer readings using the equation of Pumphrey et al. (1975):
 | (2) |
Wheat yield was regressed on the sum of soil water storage in March in the profile plus growing-season rain to evaluate the model of Leggett (1959) for this particular data set. The last neutron probe reading taken between 20 March and 10 April was used to calculate soil water storage in March. Years without readings within this interval were not used in this analysis. Rain that was received after this neutron probe reading and before harvest was summed to obtain spring rain.
Yields of pea and wheat were regressed on ET and ET/ to correct for seasonal differences in atmospheric evaporative demand (Gregory, 1984; Payne, 1997). Daily maximum and minimum relative humidity and air temperature were used to obtain using equations found in Campbell (1977).
Wheat yields were modeled using multiple regression, with winter (Oct. – Mar.), April, May, and June rains as inputs (Ramig and Pumphrey, 1977). The pea yield model of Pumphrey et al. (1979), which was developed using data from 1945 to 1977, was used to evaluate its ability to predict fresh pea yield for this independent data set. The Pumphrey model combines monthly precipitation data with an index of the heat stress during the blooming and pod-filling periods in May and June:
 | (3) |
is a constant, b1 through b5 are regression coefficients, x1 is the precipitation sum for October through March, x2 is the precipitation sum for April, x3 is the precipitation sum for May, x4 is the precipitation sum for June, and x5 is the HDDS for that particular year. The HDDS was defined by Pumphrey et al. (1979) as
 | (4) |
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsNOTESResultsDiscussionREFERENCES |
|---|
followed a remarkably conservative pattern (Fig. 1a)
. It increased from 0.1 kPa at the beginning of the year to 1.2 kPa near day of the year (DOY) 220, corresponding to late July and early August. Thereafter, values decreased rapidly until DOY 300, or early November, and continued to decrease at a lesser rate until the end of the year. The tended to be greater for pea than for wheat because pea was planted in the spring when temperatures were rising and the atmospheric humidity was falling (Fig. 1b and 1c). The exception was in 1974 when spring wheat was grown. Because the ratio of yield/ET decreases as increases (e.g., Tanner and Sinclair, 1983), the data illustrate that WUEET will almost always be greater for winter wheat than for spring wheat under these growing conditions.
|
for pea while there was perhaps a weak inverse relation between precipitation and for wheat.
|
|
|
Tenderometer scores differed due to tillage in 1968, 1970, 1971, 1976, 1978, 1979, 1981, 1983, and 1984 (Table 2). When data from all of the years were combined, tenderometer scores for the Spring MBD-MBD treatment tended to be about four points greater than those of the Fall SWP-SKW treatment. The tenderometer readings for the Fall SWP-SKW tillage treatment were slightly less than the optimum range of 90 to 110 (Pumphrey et al., 1975; Kraft et al., 1991). Because plant development is governed strongly by temperature (Olivier and Annandale, 1998), this suggests that pea in the Fall SWP-SKW treatment tended to mature more slowly than in the other treatments, perhaps due to cooler soil temperatures in the spring.
The curvilinear nature of Eq. [2], which Pumphrey et al. (1975) proposed to standardize fresh pea readings to a tenderometer score of 100, made a proportionally greater correction for tenderometer readings <95. For example, pea yield increased from 6076 to 8657 kg ha-1 in 1972 when the tenderometer reading of 89.5 was substituted into Eq. [2]. Incorporating the tenderometer readings did not alter the significance of the tillage treatments in any year although groupings of means did change for certain years, e.g., 1968 and 1977 (Table 2).
Our data (Fig. 3a) suggest that there is no unique relation between yield and tenderometer score. Trends that were particular to individual years occurred, but there was no obvious relation for other years. Indeed, some years, e.g., 1972, 1974, and 1984, the data were even inconsistent with the generalized statement that yield increases as tenderness decreases (i.e., as tenderometer readings increase). Equation [2] was developed by Pumphrey et al. (1975) using sequential harvest data, apparently during the same year, from four dryland experiments (Fig. 3b). Yield and tenderness data that were taken the same year from irrigated experiments resulted in a different empirical curve to describe the relation between yield and tenderness. Their data demonstrate what is already well known, i.e., pea becomes harder and heavier as it matures. However, our data, taken under experimental conditions that were quite different, demonstrate that Eq. [2] probably cannot be applied to independent data sets. Factors that complicate efforts at developing a single relation between yield and tenderness include wind, temperature, humidity, available soil moisture, and soil fertility (Pumphrey et al., 1975). The need to compare yields of fresh pea at a common tenderometer reading remains, but there is no consistent empirical relation between the two.
|
between wheat WUEET and yield (Fig. 4)
that was consistent with Stewart's (1989) conclusion that the greatest WUEET is achieved at the greatest yield. Figure 4 suggests that the WUEET continued to increase with yield under conditions of similar management and variable water supply because most of the points are from years during which N application was constant (45 kg N ha-1 from 1967–1981). This contrasts somewhat with Ritchie's (1983) analysis of Jensen and Sletten's (1965) data for wheat grown at a fixed N fertilization rate and variable water supply, which suggested an approximately constant value of WUEET for a given management level. Their data, which was obtained under irrigated conditions, ranged from 4 to 8 Mg ha-1 for yield and from 17 to 20 kg ha-1 mm-1 for WUEET. These represent greater values than we generally obtained under dryland conditions in the inland Pacific Northwest. Under our conditions, WUEET does not appear to have asymptotically approached an upper limit predicted by Viets' (1962) Model B for the relation between WUEET and yield.
|
|
|
Yield Models
Winter Wheat
Using Leggett's (1959) model for winter wheat, we obtained
 | (5) |
, where yield is in kilograms per hectare, and soil water storage in March and rain are in millimeters. This compares well to Leggett's original equation for winter wheat:
 | (6) |
 | (7) |
. Presumably the most important reason for this model's improvement over the Leggett model is that the latter does not take into account the fall, winter, and early spring ET.
|
further improved the model (Fig. 5c). For this data set, we obtained
 | (8) |
. This improvement illustrates the considerable influence of on crop WUE, as predicted by theory (Tanner and Sinclair, 1983). The advantage of correcting for is clearly shown for the 1974 spring wheat data, which fell below the regression line (Fig. 5b) because of a greater during its growing season (Fig. 1). The major management factor that would influence in the inland Pacific Northwest is the planting date.
Although use of the ET and ET/ offered considerable improvement over the Leggett model for wheat yield, there was still substantial scatter around the model curve. One likely source of this is winter runoff that occurs when rain falls on frozen or saturated ground (Zuzel, 1994). Runoff was not measured in this experiment and is ignored in Eq. [1]. Another contributing factor may have been the increased rates of N used during the later years of the experiment.
Using the multiple-regression approach of Ramig and Pumphrey (1977) for winter wheat, we obtained the equation
 | (9) |
. By comparison, Ramig and Pumphrey (1977) found that, on average, one gains 3.8, 14.8, and 30.7 kg ha-1 wheat for each millimeter of winter, May, and June precipitation while losing 0.1 kg ha-1 wheat for each millimeter of April precipitation. They obtained an R2 of 0.64, which is similar to our results. Overall, Eq. [9] gave a slightly better fit to the wheat yield data than the ET-based equations. However, a comparison of our results with those of Ramig and Pumphrey (1977) suggests that the coefficients were specific to the data set. For example, multiple regression using their data suggests a yield loss associated with increased April precipitation, whereas with our data set, it suggests a positive effect. Their results also suggest a greater positive effect associated with May rain than does our data set.
The residual errors of two multiple-regression models (ours and the original one of Ramig and Pumphrey, 1977) for the wheat yield data of this study are shown in Fig. 6
, along with the residual errors for the wheat ET/ model. The multiple-regression models tended to overestimate values in certain regions of the abscissa while underestimating it in others. Overall, the original model of Ramig and Pumphrey (1977) tended to underestimate the yield, especially at values near 5000 kg ha-1, and overestimate it at values near 3500 kg ha-1 (Fig. 6b). Because the ET/ model is more physically based than the regression models, it should be more robust and better suited for independent data sets. Nonetheless, the favorable results that were obtained for the regression models, in view of their simple and easily obtained inputs, render them an attractive alternative for some site-specific applications.
|
 | (10) |
. This compares reasonably well with Pumphrey's original results
 | (11) |
. The two data sets gave similar R2 values, but the values of the coefficients are different, which is similar to the multiple-regression results for wheat. In particular, their coefficients for April precipitation and HDDS were negative, whereas ours were positive, and their coefficient for May precipitation was much lower than ours.
Using linear regression for pea yield and ET, we obtained
 | (12) |
. When using ET/ instead of ET, the results were
 | (13) |
. The residual errors of these three models confirm that the Pumphrey model predicted pea yield much better than the models based on ET and ET/ (Fig. 7) , despite the change in the model coefficients. Dividing ET by actually reduced model goodness of fit, perhaps because temperature and humidity directly affected pea yield in ways that does not take into account. Possibilities for this include humidity effects on disease incidence, temperature effects on flower and pod abortion, or temperature and humidity effects on the relation between yield and tenderness. For all three models, the 1972 data deviated considerably from predicted values for unexplained reasons.
|
models is due to a combination of disease and sensitivity of pea to high temperatures per se. Indeed, this analysis suggests that water supply tends to not be the limiting factor to fresh pea production in this environment. |
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsNOTESResultsDiscussionREFERENCES |
|---|
Whether there are economic advantages to these systems is less clear, especially for the Spring MBD-MBD treatment. Given the relatively narrow window for optimum soil conditions in the spring, it would be difficult for producers with thousands of hectares to effect spring tillage operations and still plant pea in a timely manner. The fact that there has been a decline in recent years in wheat yield in the small experimental plots of the Fall SWP-SKW treatment (Payne et al., 2000), apparently due to downy brome infestation, raises the question of how easily downy brome and other weeds can be controlled in conservation systems on a farm scale. Despite markedly beneficial effects on soil physical properties, conservation tillage systems that fail to control downy brome and other pests will probably not be economically viable (Young et al., 1994; Kraft et al., 1991).
For wheat, our study suggests that using ET as a predictor for yield offers substantial improvement to the formula of Leggett (1959). An advantage of the Leggett (1959) formula is its limited data requirement. However, there are now several models available (e.g., Rickman et al., 1996) that predict wheat ET reasonably well from simple soil and weather inputs. These can be used to improve the prediction of yield potential, and therefore the estimation of N requirement, over the Leggett model. Generally, farmers cannot add fertilizer later than early March because moisture will be insufficient for the crop to take advantage of it. Therefore, both the Leggett and ET-based models require a prediction of growing-season precipitation from meteorological forecasts or historical records.
This study reaffirms the importance of in determining yield–ET relations. For this Mediterranean-type environment, winter crops will generally be at an advantage for WUEET compared with spring crops because a greater portion of the growth cycle occurs during periods of lower . Use of ET/ will probably be more reliable for use and comparison with independent data sets on yield and water use than ET alone.
The regression formula developed by Pumphrey et al. (1979) proved to be a much better predictor of pea yield that the ET-based equations. This suggests that pea yield, which is notoriously unstable in this region, is limited by other constraints such as disease and temperature extremes. However, the formula presented by Pumphrey et al. (1975) for standardizing fresh pea yields to 100% tenderometer readings appeared to be year-specific and unsuitable for general application.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsNOTESResultsDiscussionREFERENCES |
|---|
Received for publication December 8, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsNOTESResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Machado, S. Petrie, K. Rhinhart, and R. E. Ramig Tillage Effects on Water Use and Grain Yield of Winter Wheat and Green Pea in Rotation Agron. J., January 11, 2008; 100(1): 154 - 162. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Chen, P. Miller, F. Muehlbauer, K. Neill, D. Wichman, and K. McPhee Winter Pea and Lentil Response to Seeding Date and Micro- and Macro-Environments Agron. J., October 31, 2006; 98(6): 1655 - 1663. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. A. Payne, C. Chen, and D. A. Ball Agronomic Potential of Narrow-Leafed and White Lupins in the Inland Pacific Northwest Agron. J., November 1, 2004; 96(6): 1501 - 1508. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Chen, W. A. Payne, R. W. Smiley, and M. A. Stoltz Yield and Water-Use Efficiency of Eight Wheat Cultivars Planted on Seven Dates in Northeastern Oregon Agron. J., July 1, 2003; 95(4): 836 - 843. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Chen and W. A. Payne Measured and Modeled Unsaturated Hydraulic Conductivity of a Walla Walla Silt Loam Soil Sci. Soc. Am. J., September 1, 2001; 65(5): 1385 - 1391. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
