| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Pasture Systems and Watershed Management Research Unit, USDA–Agricultural Research Service, Building 3702 Curtin Road, University Park, PA 16802
b current address, AgResearch Limited, Invermay Agricultural Centre, Puddle Alley, Private Bag 50034, Mosgiel, New Zealand
c Pasture Systems and Watershed Management Research Unit, USDA–Agricultural Research Service, Building 3702 Curtin Road, University Park, PA 16802
d Crop, Soil, and Environmental Sciences, 115 Plant Science Building, Fayetteville, AK 72701
e Univ. of New Hampshire Cooperative Extension, Agricultural Resources, 3785 Dartmouth College Highway, Box 8, North Haverhill, NH 03824
f Pasture Systems and Watershed Management Research Unit, USDA–Agricultural Research Service, Building 3702 Curtin Road, University Park, PA 16802. Mention of trade names does not imply endorsement by the USDA
* Corresponding author (peter.kleinman{at}ars.usda.gov).
Received for publication January 8, 2007.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMaterials and MethodsResults and DiscussionSummary and ConclusionsREFERENCES |
|---|
Abbreviations: DRP, dissolved reactive phosphorus • NPRP, National Phosphorus Runoff Project • STP, soil test phosphorus • TP, total phosphorus
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionSummary and ConclusionsREFERENCES |
|---|
Field plots represent a spatial scale intermediate to soil runoff boxes and watersheds. Field plots of various sizes (2–622 m2) have been used to evaluate the effect of soil and management factors on runoff water quality (Pote et al., 1996; Gascho et al., 1998; Sharpley et al., 2001; Sharpley and Kleinman, 2003). As with soil runoff boxes, field plots allow for the replication of multiple management scenarios. Because they occur in situ, field plots do not destroy soil structure or alter vertical drainage characteristics as do packed soil runoff boxes. Thus, it is often assumed that field plots better represent landscape processes involved in P transport than do soil runoff boxes (Kleinman et al., 2004).
The National Phosphorus Runoff Project, or NPRP, employs the use of field plots to describe P loss in runoff (Sharpley et al., 1999). A growing number of studies have employed variations of a protocol developed under the NPRP. These studies have been aimed primarily at evaluating "source" factors within P site assessment indices such as soil P content (e.g., Pote et al., 1996), fertilizer/manure characteristics, and application method/rate/timing (e.g., Daverede et al., 2004). During NPRP simulation experiments, a constant rainfall intensity of 60 to 95 mm h–1 is applied with a rainfall simulator to generate 30 min of continuous runoff from 2-m-long plots that are approximately at field capacity at the time of rainfall initiation (Humphry et al., 2002; Daverede et al., 2004). By initiating rainfall simulations at field capacity conditions, the NPRP protocol reduces hydrologic variability between plots related to spatial differences in antecedent soil moisture as well as eliminates the need to apply large, infrequently occurring rainfall events to generate runoff. National P Runoff Project studies often report rainfall events of 5- to 15-yr return periods (e.g., Kleinman et al., 2004). By controlling initial (soil moisture) and input (rainfall) conditions, the NPRP protocol enables direct comparison of research findings across diverse rainfall-runoff experiments (e.g., Vadas et al., 2005a).
Despite successful adoption of NPRP rainfall-runoff simulation approach by many researchers to quantify source controls of P transport, it is still unclear how well these data apply to broad scales, such as hill slopes or watersheds, where management and monitoring tend to happen. Even with the apparent advantage of repeatability, field plot studies have their own uncertainties. Some properties and processes vary considerably at the field plot scale (e.g., bulk density, infiltration), such that they may contribute disproportionately to field plot observations relative to broader, landscape-scale observations. Literature reviews of field plot studies by Bloschl and Sivapalan (1995) and Wauchope and Burgoa (1995) and results from studies reported by van de Giesen et al. (2000), Chaplot and Le Bissonnais (2000), and Bagarello and Ferro (2004) point to nonlinear variations in runoff and erosional processes with increasing plot size. Therefore, the question of "upscaling" field plot findings remains a concern: exactly what processes are represented in field plots and how can quantitative findings from field plots be extrapolated to large scales?
The role of site hydrology on P transport depends largely on interactions between climate, soils, field management, and geomorphology (Gburek and Sharpley, 1998). Surface runoff mechanisms affecting P transport can be broadly distinguished into two, non-exclusive categories: infiltration-excess runoff and saturation-excess runoff. Runoff by the infiltration-excess mechanism is largely determined by rainfall intensity in excess of soil infiltration capacity and rainfall duration. Runoff by the saturation-excess mechanism occurs in response to a rising water table (perched or otherwise) that saturates the storage capacity of the surface soil. Storm characteristics are less important in determining saturation-excess runoff; even low intensity, low duration storms can produce runoff from saturated or waterlogged soils (Srinivasan et al., 2002). According to Nash et al. (2002), saturation-excess runoff includes both rainfall and soil water, while infiltration-excess runoff is predominantly rainfall water. Srinivasan (2000), in an intensive study of runoff generation from a grassed, colluvial soil in Pennsylvania showed that both the saturation and infiltration excess runoff mechanisms can occur simultaneously during a single storm. Further, Srinivasan (2000) determined that frequent, low intensity storms were predominated by saturation-excess runoff and that sporadic, high-intensity thunderstorms tended to generate infiltration-excess runoff. Related research by Needelman (2002) showed that saturation-excess runoff from cultivated soils was promoted by the presence of subsurface features, such as a fragipan or pronounced argillic horizon, that temporarily perched water, whereas runoff from soils with minimal subsurface discontinuities occurred as infiltration excess.
Given the paucity of information on hydrologic processes represented at small plot scales, particularly with regard to implications to P transport, the goal of this study was to examine the role of hydrology in plot-scale rainfall-runoff simulation studies. Major objectives of the study presented were to elucidate approaches to hydrologic interpretation of NPRP data and to evaluate plot-scale hydrology in the context of known landscape-scale hydrologic processes. Additional objectives were to evaluate the effects of soil moisture conditions, seasons, and surface management conditions (manured versus unmanured) on runoff and P transport processes.
|
Materials and Methods |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionSummary and ConclusionsREFERENCES |
|---|
Five locations in Onteora soil and seven in Marlow soil were selected for the rainfall-runoff experiments. All the locations were within 200 m from each other. Each location included two abutting plots, 2 m long and 1 m wide. Thus, ten plots were established in Onteora soil and 14 plots were established in Marlow soil. Slope gradients of these plots varied from 5 to 7%, with no significant difference between soils. For the Onteora soil, all plots had established forbs and grass swards. For the Marlow soil, all plots had established grass swards. The forbs and grasses included orchardgrass (Dactylis glomerata), timothy (Phleum pretense), tall fescue (Festuca arundinacea), perennial rye grass (Lolium perenne), red and white clover (Trifolium spp.), and alfalfa (Medicago sativa) mowed to a height of 7.5 cm. None of the plots had received manure or mineral fertilizer in the 6 mo before the experiments.
Each plot was isolated along the upper three sides by steel frames driven 5 cm into the soil and extending 5 cm above the soil. A gutter, inserted 5 cm into the soil with the upper edge level with the soil surface, at the lower end of the plot allowed routing of runoff. The gutter was equipped with a canopy to exclude direct input of rainfall. A 2-cm diameter plastic tube routed runoff from the gutter to plastic collecting vessels.
Ten soil samples were collected from a 0- to 5-cm depth adjacent to each 1 by 2 m plot and combined into a single, composite sample to represent antecedent soil P conditions at each plot. Soils were air dried, sieved (2-mm), and analyzed for Mehlich-3 P by shaking 2.5 g of soil with 25 mL of Mehlich-3 solution (0.2 mol L–1 CH3COOH + 0.25 mol L–1 NH4NO3 + 0.015 mol L–1 NH4F + 0.013 mol L–1 HNO3 + 0.001 mol L–1 EDTA) for 5 min (Mehlich, 1984). Phosphorus in the Mehlich-3 extracts was measured colorimetrically by the modified method of Murphy and Riley (1962), with 
= 712 nm.
Rainfall-Runoff Simulation Experiments
Portable rainfall simulators equipped with TeeJet 1/2 HH SS 30 WSQ nozzles (Spraying Systems Co., Wheaton, IL) approximately 3 m above the soil surface were used. During the simulations, rainfall was delivered at approximately 60 mm h–1 until 30 min of continuous runoff was collected at the outlet of the plots. Figure 1
outlines the data collection scheme during the simulation experiments. Runoff volumes were measured at 5-min intervals to generate a hydrograph for each 30-min runoff event. Runoff collected over the 30-min runoff period was thoroughly stirred to resuspend settled particles and sampled immediately. Filtered (0.45 µm) and unfiltered subsample were obtained and stored at 4°C before analysis. Dissolved reactive phosphorus (DRP) was determined on the filtered sample by the colorimetric method described above. Total P was measured on unfiltered runoff water by modified semimicro-Kjeldahl procedure following Bremner (1996). Total solids were also determined on unfiltered runoff samples by oven-drying samples at 70°C for 48 h.
|
) of the soil surface (0–6 cm), was measured using a capacitance sensor (Theta Probe, Delta-T Devices Ltd., Cambridge, UK) at six points within each plot—two at the upper end, two at the lower end, and two toward the middle part of the plot, as shown in Fig. 1 inset. The initial and final moisture contents were measured approximately at the same locations. It was hypothesized that this measured soil moisture might help to identify if a moisture gradient existed. Moisture gradient may suggest the relative propensity of areas to generate runoff, and may not represent the actual runoff generation areas within the plot. Rainfall-runoff experiments were conducted on Onteora soil plots in October (fall season) 2001 and May (spring) 2002 and on Marlow soil plots in July/August (fall) 2002. Fall season was characterized by dry conditions and spring by wet conditions. All fall simulations were conducted for three consecutive days, while spring simulations were limited to two consecutive days.
In Onteora soil, during fall, three simulations were performed on consecutive days on unmanured plots to evaluate P loss under unmanured conditions. Immediately following the third day simulation, fresh dairy manure was surface applied to the plots at an average total P addition of 75 kg ha–1. Five days following this manure application, three additional rainfall simulations were conducted on consecutive days. During spring in Onteora soil, the plots were established adjacent to locations where the initial, fall simulations were conducted, as the initial plot conditions had been modified by manure application during the fall experiments. The two-day spring simulations included rainfall-runoff experiments on unmanured and manured plots, similar to those of fall simulations. Soil samples were collected before fall and spring simulations to evaluate the soil P levels.
In Marlow soil, the first two days of simulations were conducted on unmanured soil plots. Immediately following the second-day simulation, fresh dairy manure, at a total P rate addition of 24 kg ha–1, was surface-applied to the plots. Third day simulations were conducted on freshly manured plots.
Statistical Analyses
Soil moisture, rainfall, runoff, and P transport data were evaluated using descriptive statistics, Student's t test, and general linear model combined with Tukey's mean comparison. Statements of significance in the text are based on an 
of 0.05. Relationship between P in runoff and soil test P were evaluated by least squares regression. All analyses were conducted with Minitab software, release 13.1 (Minitab Inc., 2001).
|
Results and Discussion |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionSummary and ConclusionsREFERENCES |
|---|
|
|
Effect of surface-applied fresh manure on soil moisture contents was significant. On Onteora soil, during fall season, the initial soil moisture contents on Days 2 and 3 were consistently but not significantly greater under manured than under unmanured conditions. The final moisture contents on all 3 d of the Onteora fall simulations were significantly greater for manured than unmanured conditions. However, owing to wet antecedent conditions, no significant differences in moisture contents (initial and final) were observed between manured and unmanured conditions during the spring simulations on Onteora soil. For the Marlow soil, where manure was applied after Day 2 simulations, the initial and final moisture contents recorded on Day 3 were greater, but not significantly, than Days 1 and 2, reflecting the effect of manure on soil moisture (Table 2). Neave and Abrahams (2002), from a plot-scale rainfall simulation study, observed that surface litter could lead to dams, blocking, and ponding overland flow. In the current study, such an effect would be expected to have produced consistently high final moisture contents for manured plots during the fall simulations. As the plots were at or near saturation during spring season, such effects due to manure presence on the surface might not have been evident. Also, following rainfall application, manure can release the absorbed flow, increasing the soil moisture contents of manured plots.
Rainfall-Runoff Dynamics
For a given season, for both Onteora and Marlow soils, rainfall-runoff dynamics appeared to be closely related to initial moisture conditions within the plots. Generally, as plot moisture conditions increased with successive rainfall applications, the rainfall needed to generate 30 min of continuous runoff under manured and unmanured conditions showed a decreasing trend (Table 2). Thus, Day 1 simulations required the greatest amount of rainfall to generate runoff, as antecedent moisture conditions were the driest. The depths of rainfall applied on Day 1 were significantly greater than other days of simulation for all soil, surface, and seasonal conditions tested in this study.
Within a given season, the surface application of fresh manure on Onteora soil resulted in a mixed response. Under fall conditions, the average rainfall depth was greater (not significant) under manured conditions than under unmanured conditions. Under spring conditions, the trend was reversed (compare average rainfall depth data for manured and unmanured conditions for Onteora soil, Table 2). Keeton et al. (1970) observed that dry manure packs could absorb rainfall equivalent to half of the depth of the manure. In the current study, it is likely that manure applied to the dry plots during fall season absorbed rainfall, delaying the runoff. In spring at the Onteora site, the depth of rainfall applied on Day 1 to the manured plots was significantly less than that applied to the unmanured plots, while on Day 2, there were no significant differences in rainfall applied between manured and unmanured plots. Wet site conditions during the spring likely saturated the manure, eliminating the manure-absorption effect that was observed in fall. Thus, the absence of similar trends in rainfall depths during spring and fall can be attributed to the differences in soil moisture for these two seasons.
On the Marlow soil, manure application after the Day 2 simulation did not appear to affect the depth of rainfall on Day 3 as compared to Day 2 (Table 2). Fresh manure surface applied to soil at or near saturation (e.g., Marlow soil Day 3 simulations, Onteora soil spring simulations) likely did not absorb as much moisture as manure applied under dry soil conditions (e.g., Onteora fall simulations).
While the depth of applied rainfall declined with successive simulations, the depth of runoff tended to increase with successive rain simulation events, with the largest increases occurring from Day 1 to Day 2. For the Onteora plots, for both spring and fall seasons and for both surface conditions, the increasing depth of runoff with successive events was not statistically significant. For the Marlow soil plots, the depth of runoff measured on Day 1 was significantly less than those measured on Days 2 and 3, and there were no significant differences in runoff depths between Days 2 and 3. The absence of significant differences in initial moisture contents, depths of rainfall applied to generate runoff, and runoff generated on Days 2 and 3 for both soils suggests that the field plots were at or near equilibrium in converting rainfall to runoff by Day 2.
To gain further insight into the conversion of rainfall to runoff and to characterize the runoff generation process, the runoff-rainfall ratio, or RR ratio, during the runoff generation periods was calculated (Fig. 2 and 3 ). Before the initiation of runoff, the RR ratio is zero, signifying that all of the applied rainfall is stored (combination of infiltration and surface storage) within the plots. As runoff initiates, the applied rainfall is partitioned into runoff and plot storage. Surface retention of rainfall can be neglected as it is limited, and, usually, is fulfilled within the first few minutes of rainfall application. Under circumstances where infiltration continues even after the initiation of runoff, the resulting runoff can be viewed as infiltration-excess, where rainfall intensity and intake capacity of soil control runoff generation. Thus, for infiltration-excess runoff, the RR ratio remains less than 1. In contrast, when the soil is completely saturated with no capacity to infiltrate input rainfall, the applied rainfall is converted directly to saturation-excess runoff. Thus, the RR ratio for saturation-excess runoff is equal to 1.
|
|
On Onteora soil, the RR ratios were consistently and significantly greater during the spring simulations than the fall simulations under all surface conditions (Fig. 2). This relates very well with the high runoff depths recorded during spring periods (Table 2). Despite diverse soil moisture conditions, the average depths of rainfall were not significantly different between spring and fall seasons for the Onteora soils. Thus, the wet soil conditions in spring did not alter the time to runoff and the runoff generation mechanism (infiltration excess) but resulted in greater conversion of rainfall into runoff (higher RR) and generated more runoff in spring than in fall.
Based on the results from Onteora soil plots for spring and fall seasons, it appears that rainfall simulations on saturated soils can result in a larger conversion of rainfall into runoff and greater runoff depths than those observed from soils in field capacity conditions. A high RR ratio and a large runoff depth was indicative of large available (sediment and nutrient) transport capacity in spring. During fall simulations on the Onteora soil, as initial soil moisture conditions increased with successive days of rainfall applications, RR ratios increased, though not close to those under spring conditions. However, unlike fall, during spring, the RR ratio did not appear to vary with initial conditions, as evidenced by a lack of difference in RR ratios at the start of runoff on Days 1 and 2 of spring (refer to Fig. 2, 0–5 min RR ratios). In the case of the Marlow soil, the RR ratios were consistently increasing with successive simulations. Even though there were no significant differences in runoff depths between Days 2 and 3 from the Marlow plots, the RR ratios were consistently but not significantly greater on Day 3 than on Day 2.
Surface application of manure did not significantly affect runoff depths for both Onteora and Marlow soils (Table 2). In addition, no significant differences in RR ratios were observed between manured and unmanured plots (Fig. 2 and 3). This suggests that storage of water by surface-applied manure, while significant from the standpoint of delaying the timing of runoff, does not change the infiltration capacity of the soils after runoff is initiated.
Phosphorus Transport
Before manure application, DRP averaged 6 to 40% of TP in runoff from the Onteora soil, with a maximum daily average of 11% under dry, fall conditions and 40% under wet, spring conditions (Table 3
). In the case of Marlow soil, DRP averaged 63 to 69% of TP in runoff before manure application. As illustrated with fall and spring Onteora data in Fig. 4
, sediment in runoff influenced the relative contribution of DRP to TP in runoff. Specifically, the ratio of DRP to TP in runoff was weakly negatively related to the total solids concentration in runoff. Sharpley et al. (1981) identified sorption of dissolved P by suspended sediment as a key process controlling dissolved P concentrations in runoff. Following manure application, DRP became the dominant form of P in runoff from all soils (averaging 85% for the Onteora and 71% for the Marlow) as water-soluble P in the manure likely served as the primary source of P in runoff.
|
|
|
Hydrological Considerations
In the northeastern U.S., saturation excess is increasingly viewed as a dominant mechanism of runoff in settings where vertical discontinuities in soil properties ("shallow soils") coincide with lower landscape positions where lateral flow can accumulate leading to saturated soil conditions (Gburek et al., 2006). Walter et al. (2003) considered saturation-excess runoff to be the principal form of runoff in the regions of New York, where the present study was conducted. However, rainfall application using rainfall simulators on Onteora soil resulted in infiltration-excess runoff for both dry fall and wet spring seasons. Hence, the relationships derived based on these plot studies might only be applicable to similar field conditions. Ward (1984) indicated that saturation-excess runoff is a landscape-scale process and that infiltration-excess is controlled by local soil properties. In the case of plot-scale studies, as the one described here, local soil properties may play a major role in controlling the hydrologic responses. Also, application of rainfall, using a rainfall simulator, over a limited area is unlikely to influence landscape hydrologic process. To represent saturation-excess runoff conditions using small plots, the plot may need to be located in areas saturated due to landscape-scale processes such as the presence of a shallow water table or subsurface lateral flows from upslope areas.
|
Summary and Conclusions |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionSummary and ConclusionsREFERENCES |
|---|
This study examined the hydrology of small plots established in two soils with contrasting properties and subjected to a modified version of the NPRP protocol. For both soils, absence of significant moisture gradients within the plots before and after rainfall simulations indicated the uniformity of runoff generation. Application of manure to plots under dry antecedent conditions delayed overland flow initiation and increased soil moisture by enhancing infiltration. Application of manure to plots at field capacity or near-saturation conditions did not affect rainfall depth or runoff initiation. For the Onteora soil, application of manure in the spring created faster initiation of runoff than was observed from plots before manure application.
National P Runoff Project experiments, when conducted under wet climatic conditions, such as springtime in the present study, needed more than 24-h drainage period between successive simulations. Soil moisture conditions during wet, spring periods indicated that the plots did not sufficiently drain to field capacity conditions on successive days of simulations. Generally, wet antecedent conditions during spring might have resulted in saturation-excess runoff as opposed to dry fall conditions, which resulted in infiltration excess runoff. However, with rainfall simulators at plot scale, saturation-excess runoff could not be generated. While developing soil P-runoff P relationships, it is necessary to represent the runoff generation processes, as the rainfall return periods resulting in runoff generation processes can be widely different. Even though the wet spring season did not alter the runoff generation mechanism (infiltration-excess) as those observed during dry fall season, a larger conversion of rainfall to runoff and greater runoff depths were observed in spring than in fall. This could significantly increase the transport capacity in spring as opposed to fall.
Spring simulations resulted in significantly larger dissolved P loss than fall, though the total P concentration between those two seasons was not significantly different. Freezing and thawing cycles during winter periods appeared to have resulted in large availability of DRP for transport during spring periods. Upon manure application, both spring and fall seasons resulted in dissolved P loss that constituted 80 to 90% of total P loss. In the case of Marlow soil, sandy soil structure resulted in high (60% of TP was DRP) and constant dissolved P losses under manured and unmanured conditions.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionSummary and ConclusionsREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionSummary and ConclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. W. McDowell, D. Nash, A. George, Q. J. Wang, and R. Duncan Approaches for Quantifying and Managing Diffuse Phosphorus Exports at the Farm/Small Catchment Scale J. Environ. Qual., August 24, 2009; 38(5): 1968 - 1980. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
