Published in J. Environ. Qual. 33:192-200 (2004).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA
TECHNICAL REPORTS
Landscape and Watershed Processes
Dynamics of Potassium Leaching on a Hillslope Grassland Soil
Marta A. Alfaro*,a,
Peter J. Gregoryb and
Stephen C. Jarvisc
a National Institute for Agricultural Research (INIA), Remehue Research Station, Casilla 24-O, Osorno, Chile
b The University of Reading, Department of Soil Science, Whiteknights, PO Box 233, Reading, Berks, RG6 6DW, UK
c Institute of Grassland and Environmental Research (IGER), North Wyke Research Station, Okehampton, Devon, EX20 2SB, UK
* Corresponding author (malfaro{at}remehue.inia.cl).
Received for publication October 2, 2002.
 |
ABSTRACT
|
|---|
There have been only a few studies of potassium (K) losses from grassland systems, and little is known about their dynamics, especially in relation to nitrogen (N) management. A study was performed during the autumn and winter of 1999 and 2000 to understand the effects of N and drainage on the dynamics of K leaching on a hillslope grassland soil in southwestern England. Two N application rates were studied (0 and 280 kg N ha–1 yr–1), both with and without tile drainage. Treatments receiving N also received farmyard manure (FM). Higher total K losses and K concentrations in the leachates were found in the N + FM treatments (150 and 185% higher than in 0 N treatments), which were related to K additions in the FM. Drainage reduced K losses by 35% because of an increase in dry matter production and a reduction in overland and preferential flow. The pattern of change in K concentration in the leachates was associated with preferential flow at the beginning of the drainage season and with matrix flow later in winter, and was best described by a double exponential curve. Rainfall intensity and the autumn application of FM were the main determinants of K losses by leaching. The study provided new insights into the relationships between soil hydrology, rainfall, and K leaching and its implications for grassland systems.
Abbreviations: D, drainage treatment effect • D+, drained plot • D–, undrained plot • FM, farmyard manure • MTF, mole plus tile drain flow • N, nitrogen treatment effect • N+, 280 kg N ha–1 yr–1 • N–, 0 kg N ha–1 yr–1 • OSF, overland plus subsurface lateral flow •
|
INTRODUCTION
|
|---|
LITTLE ATTENTION has been given to potassium (K) leaching because this element does not result directly in eutrophication. However, there have been many previous agronomic studies that have shown that, because K is a mobile ion in soils, significant amounts can be lost by leaching (Quemener, 1986) thereby affecting the efficiency of the fertilizers applied. The broad range of factors affecting total K losses are well known. Potassium losses by leaching are influenced by the amount of K applied as fertilizer, the crop, the type of soil, and the amount of drainage (Johnston and Goulding, 1992). Nevertheless, little work has been conducted to understand the dynamics of K leaching and the effects of soil hydrology on the pattern of K losses.
The recovery of K applied in fertilizers is greater in grass crops than in arable cropping systems and can be as high as 90% of the total K applied (Pearson and Ison, 1997). This efficiency increases with the use of nitrogen (N) fertilizer and with best management practices. Bolton et al. (1970) showed that N fertilizer increased plant uptake of K and, because of this, reduced K losses by leaching by approximately 20%. In southwestern England, soil K balances were strongly affected by the dry matter production of the pasture and losses were apparently related to the amount of available soil K, which in turn was related to plant uptake (Alfaro et al., 2003).
In the UK, grassland soils show a wide range of drainage properties and differences in the magnitude of K leaching losses between undrained and drained soils are not clear (Alfaro, 2002). Because of the increase in grass growth in drained areas compared with undrained areas, a reduction of K in soil solution would be expected, which might reduce the amount susceptible to leaching.
The objective of the present study was to establish the effects of N application and drainage on K leaching losses in a hillslope grassland system on a poorly drained silty clay soil and to examine in detail the temporal pattern of losses through two drainage seasons. The hypothesis was that lower K losses would occur with high rather than low N input because this would result in less soil-available K at the beginning of the drainage season. This effect should also be greater in drained areas, where greater dry matter production is obtained through an improvement in soil moisture status. It was also proposed that K leaching would be affected by the amount of drainage flow.
|
MATERIALS AND METHODS
|
|---|
Experimental Site
Losses of K were measured on an existing field experiment at the Institute of Grassland and Environmental Research (IGER), North Wyke Research Station (southwestern England; 50°45' N, 3°50' W; 185 m above mean sea level) during the autumn and winter of 1999–2000 and the autumn and winter of 2000–2001. The soil at the site has an impermeable subsoil that results in poor drainage and is classified as a Typic Haplaquept (USDA Soil Conservation Service, 1975), Hallsworth series (Harrod, 1981). The main components of the clay fraction in the first 10 cm are illite (42%) and kaolinite (40%).
The experiment was initially established in 1982 to investigate the effects of N and drainage on livestock production (Tyson et al., 1993) and consists of 14 plots of 5 to 10% slope (Scholefield et al., 1993). Each plot was 1 ha and hydrologically isolated from its neighbors except for deep seepage, which was negligible for a subsoil with such a low hydraulic conductivity (Armstrong and Garwood, 1991). Two drainage treatments were available at the site: undrained (D–) and drained (D+) plots. The installation of mole and tile drains on the D+ plots was described by Armstrong and Garwood (1991). On the D– plots, the main hydrological pathways were overland plus subsurface lateral flow (OSF) while on the D+ plots these pathways were less important and replaced by flow through the mole plus tile drains (MTF). Details of the flow pathways can be found in Haygarth et al. (1998). It is important to note that because of the experimental layout it was not possible to separate overland from subsurface lateral flow nor was it possible to separate mole from tile flow.
The original treatments were modified in 1993 to investigate the effects of N fertilizer on environmental aspects of livestock production. A summary of the four unreplicated treatments used in the present study is provided in Table 1. Plots receiving 0 N addition (N–) were grazed by cattle but those receiving N fertilizer (N+) (280 kg N ha–1 yr–1) were subdivided in 1998 into 10 smaller paddocks comprising cut and grazed areas. The N– plots received 50 kg K ha–1 yr–1 while the N+ plots received a supplement of K fertilizer on the cut areas providing in total 150 kg K ha–1 yr–1 for these sectors. Grazed areas of the N+ treatments received 50 kg K ha–1 yr–1 and recycled K from excreta. The first N and K application was performed the first week of March each year and subsequent N additions were performed monthly after that. The supplementary K added to the N+ plots was applied after each silage cut in equal doses. The N+ plots also received applications of 14 Mg ha–1 (28 Sept. 1999) and 20 Mg ha–1 (21 Sept. 2000) of farmyard manure (FM). In 1999, N and K concentrations in the FM were 3.6 ± 0.16 and 5.1 ± 0.96%, respectively, on a dry matter basis, with a dry matter content of 16.9 ± 0.80%. In 2000, the corresponding values were 3.3 ± 0.74, 5.9 ± 0.46, and 25.5 ± 0.40%, respectively. Total N and K additions in FM were 85 and 122 kg ha–1 in 1999 and 168 and 304 kg ha–1 in 2000, respectively. All plots received the same amount of P fertilizer (Table 1).
Potassium Leaching
During autumn and winter (September to April each year), duplicate water samples were collected daily from the V-notch weir at the drainage outlet of each plot in polyethylene bottles (50 mL), which were rinsed twice with sample before the final collection. The samples collected represented drainage from either the 0- to 30-cm depth (OSF) in the D– plots or the 0- to 30- (OSF) and 0- to 85-cm depths (MTF) in the D+ plots. Because they were free of other materials, no filtration or centrifugation was necessary. Samples were kept in a refrigerator at 5 to 7°C until analysis for K by flame photometry. The amount of drainage through each weir was recorded with clockwork hydrographs. Because these were monthly charts, daily flow data were calculated as the average of the highest and lowest peak flow for that day and the total for the period was estimated as the addition of all the daily-calculated data. Sampling during periods of background flow was also considered, but these results could not always be represented graphically (flow of <0.01 L s–1) in the Results section. Also, in some cases, flow samples were the result of high water tables and spring waters, in absence of rainfall. Total K losses were calculated as the product of flow and K concentration in the samples at daily intervals and then daily loads were summed for the period.
Rainfall was measured every two minutes by an automatic weather station placed at the experimental site and summed to give the 30 min totals, which were summed to give the total daily rainfall used in this paper. Rainfall intensity was also registered automatically and the data expressed as mm rainfall h–1.
For the study of plot hydrology, flow discharges during storm events were selected from the flow discharge charts for different times of the year (dry soil, saturated soil). These were analyzed hourly for their flow distribution and the data logarithmically transformed for analysis.
Statistical Analyses
The hydrological patterns in the treatments were examined by comparing the slopes of the linear regressions obtained after the hourly flow data were logarithmically transformed. Analysis of variance (ANOVA) was conducted to establish differences in total K losses and K concentration of drainage waters as affected by the management factors (N, D). Because of the lack of replicates, this analysis assumed no interaction between years and treatments. Regression analyses were also made to establish the relationships that best described the pattern of changes in K concentration in the leachates as drainage proceeded. In all cases, Genstat 4.2 for Windows (VSN International, 2000) was used as the statistical package.
|
RESULTS
|
|---|
Rainfall and Hydrology
Although the total amount of rainfall for both drainage periods was similar (i.e., 1231 and 1266 mm for 1999–2000 and 2000–2001, respectively), its distribution through the year was different. Greater amounts of rainfall were registered at the beginning of the drainage period in 2000–2001 (3 times greater in October and 2.5 times greater in November) compared with 1999–2000. Drainage responded rapidly to rainfall in the experimental plots during 1999–2000 and 2000–2001 (Fig. 1)
, in agreement with the observations of Armstrong and Garwood (1991). Total flows were similar in plots with the same drainage treatment (Table 2), but greater totals were measured during 2000–2001, probably in response to the more intensive rainfall during autumn. In drained plots, the main pathway was MTF (69% of the total drainage on average, for both years). Average K concentration in rainfall over the experimental period was on average 0.60 ± 0.014 mg L–1.
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 1. (a) Rainfall (mm d–1), (b) flow (L s–1) in overland plus subsurface lateral flow (OSF) of undrained (D–) treatments, (c) flow (L s–1) in overland plus subsurface lateral flow (OSF) and mole plus tile flow (MTF) of drained (D+) treatments and K concentration in the leachates (mg L–1) with and without nitrogen + farmyard manure (FM) fertilizer application for 2000–2001 in (d) undrained (D–) and (e) drained (D+) treatments. Arrows indicate the date of FM application.
|
|
The flow peaks measured for both drained and undrained plots were greatest during the autumn and winter of 2000–2001. In 1999–2000, similar flow rates were measured in the OSF and MTF of D+ treatments (1.8 and 1.9 L s–1 on average, respectively). During 2000–2001, the average for OSF was 38% greater than that for MTF (3.6 and 2.6 L s–1, respectively). In the D– treatments the mean flow peaks (OSF) were 2.7 and 4.3 L s–1 for 1999–2000 and 2000–2001, respectively.
The dynamics of flow during storm events were compared using data at the beginning of the drainage period (September 1999) and during the middle of winter (February 2000) (Fig. 2)
. These two periods were chosen because the soil had different moisture contents and high proportions of the variance were accounted for (>86%) when logarithmically transformed data were used. In September 1999 the flow peaks were 3.1 and 3.5 L s–1 and 1.1 and 1.2 L s–1 for N– and N+ plots of undrained and drained treatments, respectively. The maximum rainfall registered during this event was 4.7 mm h–1. In February 2000, the flow peaks were 1.8 and 1.7 L s–1 and 0.9 and 1.0 L s–1 for N– and N+ plots of undrained and drained treatments, respectively, and the maximum rainfall intensity was 1.6 mm h–1. On both occasions, larger intercepts were found in the hydrographs of undrained plots than those of drained plots (Fig. 2). The number of hours with drainage flow was greater in the undrained plots than in drained plots in September 1999 (insertions, Fig. 2), but not in February 2000. The hydrographs also had steeper slopes in September 1999 than in February 2000 (p 
0.05; Fig. 2). Statistical analysis of the slopes of the transformed flow data showed similar intercepts and slopes (p > 0.05) for both N– and N+ plots within drained and undrained treatments. This indicates that there were no major differences between treatments in their hydrological behavior, other than the presence or absence of drainage.
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 2. Typical hydrographs during the declining phase in drainage from (a) undrained and (b) drained treatments with and without nitrogen application. Results are for a storm event during September 1999 (top) and February 2000 (bottom). The inserts show the corresponding linear regression for the logarithmically transformed flow data.
|
|
Potassium Concentration in Leachates
Similar patterns of K concentration in the leachates over time were observed during both years, but no significant effect of the treatments on the average of K concentration in leachates (p > 0.05) was found (Table 3). Higher K concentrations were obtained in N+ than in N– plots and in D– than in D+ plots during both years (p 0.05). Also, higher concentrations were found in 2000–2001 than in 1999–2000 (p 0.05; Table 3). Potassium concentrations in the OSF component from D+ treatments were not included in the calculations because this flow occurred only on a few occasions (although K concentration in the OSF samples was higher than in the corresponding MTF).
View this table:
[in this window]
[in a new window]
|
Table 3. Total K losses and K concentrations in mole plus tile flow drainage (MTF) of drained plots and overland plus subsurface lateral flow (OSF) of undrained plots from each treatment and effects of nitrogen (N) and drainage (D) for 1999–2000, 2000–2001, and the average for both years (±standard error).
|
|
The best model to describe changes in K concentration in leachates over time was a double exponential equation, which described two different rates, the first being faster than the second (Fig. 3)
. All the equations describing the change in K concentration in different treatments were highly significant (p 0.001) and accounted in most cases for large proportions (>70%) of the variance in K concentration. Equations for the same treatment were significantly different in the two years (p 0.001; Table 4) and, for each year, there were significant differences between treatments. During 1999–2000 and 2000–2001, the greatest intercept was found in the N+, D– treatment. The fastest initial rates of concentration decrease were obtained in the N+, D– treatment and the slowest in the N+, D+ treatment (p 0.01).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3. An example of a fitted double exponential curve to describe the trend of changes in K concentration (mg L–1) in drainage over time (accumulated number of days with drainage) from the N+, D– treatment during 1999–2000. Day 0 represents the concentration in the background flow at the time when the drainage started.
|
|
View this table:
[in this window]
[in a new window]
|
Table 4. Equations describing the trends in changes in K concentrations in drainage from grassland systems over time (accumulated number of days with drainage) for all treatments.
|
|
Total Potassium Losses
Total K losses ranged from 5 to 31 kg K ha–1 yr–1 depending on the year and treatments. Greater total K losses were observed in 2000–2001 than in 1999–2000 (p 0.05; Table 3). Losses were greater in N+ treatments (150% higher than from the N– treatments on average for both years; p 0.05) and D– treatments (drainage reduced the amount of K lost during both seasons by 35%; p > 0.05). No significant interactions were found between N and D (p > 0.05).
|
DISCUSSION
|
|---|
Hydrology
Hydrographs display the properties of drain flow with respect to time and show the rate of flow at all points in time during and after a rainfall or snowmelt event (Viessman and Lewis, 1996). The intercept of the hydrographs is related to the amount of rainfall and the soil storage capacity, and the slope of the recession section is related to hydraulic conductivity. In the present study, analysis of the hydrographs showed that although this experiment had unreplicated plots, those with the same drainage system had similar slopes and intercepts and, therefore, behaved hydrologically in the same way. Therefore, it can be assumed that the differences in the total amount of K lost by leaching in the different systems can be attributed to differences in the treatments applied and not to inherent differences in hydrological pathways in different plots. The results also indicate that 1-ha plots were sufficiently large to overcome the spatial variability of soil hydrology.
Greater intercepts were found in hydrographs of undrained plots compared with those from drained plots during 1999–2000 and 2000–2001. This was probably because of the shallower (0–30 cm) layer of soil from which OSF originates, which resulted in a smaller volume of soil with drainage water, so that less water was retained in the undrained than in the drained plots. Two examples of the hydrology of the plots are the results of September 1999 and February 2000. In September 1999, the water drained more quickly from undrained than from drained plots, but later in the winter (February 2000), this situation was reversed. It has been suggested before that macropore flow (OSF in the present conditions) can be an important pathway for water movement in this soil series during the rewetting period (Haygarth et al., 2000). In February 2000, there was a longer period of drainage flow in the drained plots so that in winter, matrix flow was probably the main means of water movement.
Potassium Losses
Higher K concentrations in the leachates and higher total K losses were obtained during 2000–2001 than 1999–2000. This could be for two reasons. First, in 2000–2001 greater amounts of K were applied in FM, and this was also applied the day before the first significant rainfall event of the drainage season. Second, the intensity of rainfall was higher during 2000–2001. The occurrence of higher rainfall intensities at the beginning of the autumn, when preferential flow was a major pathway of K movement, could explain the greater K concentrations in the leachates than in the rest of the season, because it would have resulted in greater amounts of K being lost directly or soon after the application of the fertilizer.
An examination of the changes in K concentration in the leachates over time showed two clear phases: (i) a very rapid decrease after the peak was reached and (ii) a period of slow decline. The rapid increase–decrease phase was always observed at the beginning of the autumn, as also measured by Hatch et al. (1997). The magnitude of this increase in K concentration was greater in the N+ treatments and was observed immediately after FM application. When FM was applied just before the first significant rainfall event, the peak in K concentration was reached the day after the application. In all cases, final concentrations (i.e., the average for the last week of sampling) were smaller than those in leachates at the beginning of the drainage season (i.e., days before the peak concentration occurred). The slopes of the equations describing the pattern of K loss were greater in the N+ than in the N– treatments during both years.
Although the total amounts of K lost were low, the average K concentrations in the leachates were greater than those reported in the literature for other grassland (Heng et al., 1991; Bache, 1990) and cropping (Bache, 1990; Bolton et al., 1970) systems. This is probably a consequence of the greater amounts of K applied as fertilizer in the present study, especially in the N+ treatments.
Leaching losses were greater in the N+ treatments, almost certainly because of the application of FM. Edwards and Daniel (1993), in an experiment with pig slurry applied to grass swards, showed that losses of nutrients, including K, were large during storms if they occurred soon (24 h) after application, as happened during the present study. Also, at the end of the growing season there was less available soil K in N+ treatments (Alfaro, 2002), and lower concentrations were measured in the leachates from these areas when the drainage season started. Higher K concentrations in the leachates were found in the N– treatments, coincident with greater amounts of plant-available K at the end of the growing season (Alfaro et al., 2003). On average of both years, N+ treatments had 25% less soluble soil K than the N– treatments (18 ± 2.4 and 24 ± 3.1 mg kg–1, respectively; p 0.01) and 43% less exchangeable soil K than the N– treatments (99 ± 10.9 and 173 ± 27.1 mg kg–1, respectively; p 0.05). These results support the notion that the exhaustion of K from the soil (and with this, its availability for leaching) can result from increases in yield and K concentration in plant tissues given high N fertilizer applications (Thelier-Huche et al., 1999). Similar effects were observed with artificial drainage. In D+ treatments there was greater dry matter production during the spring and summer (Alfaro, 2002) and thus less K in soil solution available to be leached.
Potassium Leaching Dynamics
The present results and those of Ulen (1999) suggest that K leaching losses have two components: (i) a rapid loss, induced mainly in the topsoil, which is transported by preferential flow, and (ii) a slower loss through matrix flow, shown as Phases A and B in Fig. 4
. In undrained areas, during winter, only OSF was observed and this reached higher peaks and lasted for fewer hours than the MTF of drained areas. It is probable that at the beginning of a rain event, and depending on the intensity of the rainfall, the main component of the OSF will be overland and preferential flow in the upper 30 cm of the soil, through cracks and fissures in the soil and channels left for old roots and earthworms, with a rapid increase and decrease in K concentration in leachates (Phase A). Once the soil has reached field capacity, the development of matrix flow will generate a much slower rate of K loss (Phase B). If there is no K input before the first significant rainfall event (i.e., the one with a greater effect on macropore flow), both intercept and rate of K losses are likely to be lower in these treatments.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4. Idealized hydrographs showing (a) overland plus subsurface lateral flow (OSF) and mole plus tile flow (MTF) and (b) simplified diagram of the trend in K concentration in drainage water with and without N + farmyard manure (FM) application. The Phase A and B sections indicate fast and slow phases of the pattern of K losses by leaching.
|
|
Phase A (Fig. 4) results from the presence of K+ in solution from several processes such as the release of K to the soil solution from rewetting processes, biological activity, and fresh organic matter decomposition. This release of K is followed by quick export from the system through macropore or preferential flow at the beginning of drainage period. Haygarth et al. (2000) suggested that in the present soil series (Hallsworth) during summer, soil moisture deficits result in vertical fissures, which permit vertical macropore flow and that, after rewetting, the soil reverts to an impermeable state, with only matrix flow. The hydrological importance of macropores varies over time in this soil (Haygarth et al., 2000) and because of this, changes in K concentration in drainage can also be expected over time, in agreement with the hydrological changes noted above.
Phase B (Fig. 4) of loss (determined by matrix flow) represents the movement of K in water that has moved through the profile and has reached an equilibrium with the soil solution. This explains why K concentration in the MTF of D+ treatments was always lower than that in comparable OSF samples. Matrix flow has longer to reach equilibrium with soil solution and, because it also moves through soil layers of lower K availability, this equilibrium was reached at lower concentrations, in agreement with Timmons et al. (1977). Once Phase B has been reached, and if no inputs of K occur, no other high fluxes of K will be observed until the initiation of a new hydrological cycle.
Potassium leaching losses can also be divided into background (i.e., those observed during normal rainfall periods) and occasional (i.e., resulting from the application of fertilizers or manures coincidentally with rainfall or the occurrence of high-intensity storms losses). Occasional losses were especially important during 2000–2001 because of a number of storm events. Storm events are defined as occurring when rainfall is equal to or greater than 15 mm h–1 (Boardman and Robinson, 1985). Overall loss depends upon the rate of loss and the duration of the event so that similar K losses were obtained from short but intense events and from less intensive but longer events.
In situations where the background losses are more important, higher K concentrations will be found in MTF or other deep water movement than in OSF, and the total amount of leached K will be dependent on the total amount of flow through the system. In areas where rainfall is less frequent but very intense (i.e., a predominance of storms), K concentration will be greater in OSF than in MTF and the rainfall intensity and the duration of the event will be the main factors controlling total K losses. This effect will be greater with snowmelt or on frozen surfaces with reduced water infiltration capacity (Hawkins and Scholefield, 1996; Timmons et al., 1977). Potassium can also be lost by soil erosion (Zobisch et al., 1995) or transport of soil or FM particles associated with the high turbulence and erosive capacity of OSF (Haygarth et al., 1998). However, other studies at this site (Alfaro, 2002) showed that only minor amounts (<6%) of the total K in leachates were associated with particulate materials.
During the development of specific storm events, the increase in flow resulted in a reduction in K concentration in the leachates (Alfaro, 2002). Similar behavior has been observed in the pattern of response in streams, where K concentration is affected by the rate of drainage discharge (Tsirkunov et al., 1992). When there was no increase in the discharge, the concentration remained constant (Probst et al., 1995). The only other way to modify the pattern is through the application of fertilizers or manures. Tsirkunov et al. (1992) showed that the application of FM over a wide area generated a change in the K concentration in the streams receiving the discharge. Studies by Burt et al. (1996) of the river Gara (Devon, UK) showed that K concentration increased during October, which may have been caused by surface and near subsurface runoff being relatively rich in K. Later in winter, this runoff may have been diluted by deeper throughflow of lower K concentration, which is in agreement with the present findings.
|
CONCLUSIONS
|
|---|
A field lysimeter experiment was performed to study K leaching dynamics under different nitrogen and drainage treatments on a hillslope grassland soil in southwestern England. Rainfall was the factor that most strongly affected K leaching, not only in relation to the amounts of K lost, but also with respect to the pathway of loss. Potassium concentration in the leachates over time showed two phases, first a rapid initial decrease connected with preferential flow at the beginning of the drainage period and second, a slower phase later in winter, dominated by matrix flow. Inputs of K at the beginning of the drainage period increased the intercept and the slope of the equation that described K losses over time because of an increase of available K in soil solution. Mole plus tile drainage reduced the rate of losses because of an increase in dry matter production and K uptake during the previous growing season and a reduction in the impact of OSF on leaching. Greater total K losses were found in the treatments that received 280 kg N ha–1 yr–1 + FM and in those without mole plus tile drainage. The addition of N in the system reduced the amount of K lost by leaching when no K inputs were added during the drainage period.
To reduce K leaching losses, applications of manures or mineral fertilizers should be avoided before heavy rainfall. The application of FM at the beginning of the drainage season at rates as high as those used for this experiment are not recommended. Smaller rates in a number of applications are recommended to avoid K saturation of the soil and to match fertilizer applications with plant requirements over the growing season. The application of manures after each silage cut, in replacement of mineral fertilizer, is a good option.
|
ACKNOWLEDGMENTS
|
|---|
We thank Dan Dhanoa for his advice in statistical matters, Andrew Bristow for his help in laboratory methods, and other colleagues at North Wyke for their contribution to this work. The Institute of Grassland and Environmental Research (IGER) is supported by the Biotechnology and Biological Sciences Research Council (BBSRC). Thanks to the Chilean Government and the National Institute for Agricultural Research (INIA) for funding M.A. Alfaro, and also to the Potash Development Association (PDA) for its contribution to the funding of this work.
|
REFERENCES
|
|---|
- Alfaro, M.A. 2002. Potassium balances in grassland systems. Does nitrogen affect potassium cycling and leaching? Ph.D. diss. The Univ. of Reading, Reading, UK.
- Alfaro, M.A., S.C. Jarvis, and P.J. Gregory. 2003. Potassium budgets in grassland systems as affected by nitrogen and drainage. Soil Use Manage. 19:89–95.
- Armstrong, A.C., and E.A. Garwood. 1991. Hydrological consequences of artificial drainage of grassland. Hydrol. Processes 5:157–174.
- Bache, B.W. 1990. Solute transport in soils. p. 437–462. In M.G. Anderson and T.P. Burt (ed.) Process studies in hillslope hydrology. John Wiley & Sons, Chichester, UK.
- Boardman, J., and D.A. Robinson. 1985. Soil erosion, climatic vagary and agricultural change on the Downs around Lewes and Brighton, autumn 1982. Appl. Geogr. 5:243–258.
- Bolton, E.F., J.W. Clylesworth, and F.R. Hore. 1970. Nutrient losses through tile drains under three cropping systems and two fertility levels on a Brookston clay soil. Can. J. Soil Sci. 50:275–279.
- Burt, T.P., A.L. Heathwaite, and P.J. Johnes. 1996. Stream water quality and nutrient export in the Slapton catchment. Field Stud. 8:613–627.
- Edwards, D.R., and T.C. Daniel. 1993. Runoff quality impacts of swine manure applied to fescue plots. Trans. ASAE 36:81–86.
- Harrod, T.R. 1981. The soils of North Wyke and Rowden. Soil Survey of England and Wales. Grassland Res. Inst., Harpenden, UK.
- Hatch, D.J., S.C. Jarvis, A.J. Rook, and A.W. Bristow. 1997. Ionic contents of leachate from grassland soils: A comparison between ceramic suction cup samples and drainage. Soil Use Manage. 13:68–74.
- Hawkins, J.M.B., and D. Scholefield. 1996. Molybdate-reactive phosphorus losses in surface and drainage waters from permanent grassland. J. Environ. Qual. 25:727–732.[Abstract/Free Full Text]
- Haygarth, P.M., A.L. Heathwaite, S.C. Jarvis, and T.R. Harrod. 2000. Hydrological factors for phosphorus transfer from agricultural soils. Adv. Agron. 69:153–178.
- Haygarth, P.M., L. Hepworth, and S.C. Jarvis. 1998. Forms of phosphorus transfer in hydrological pathways from soil under grazed grassland. Eur. J. Soil Sci. 49:65–72.
- Heng, L.K., R.E. White, N.S. Bolan, and D.R. Scotter. 1991. Leaching losses of major nutrients from a mole drained soil under pasture. N. Z. J. Agric. Res. 34:325–334.
- Johnston, A.E., and K.W.T. Goulding. 1992. Potassium concentrations in surface and groundwaters and the loss of potassium in relation to land use. p. 135–158. In Potassium in ecosystems—Biogeochemical fluxes of cations in agro- and forest-systems. 23rd Colloquium of the Int. Potash Inst., Prague. 17–21 Aug. 1992. Int. Potash Inst., Bern, Switzerland.
- Pearson, C.J., and R.L. Ison. 1997. Agronomy of grassland systems. 1st ed. Cambridge Univ. Press, Cambridge, UK.
- Probst, A., B. Fritz, and D. Viville. 1995. Mid-term trends in acid precipitation, streamwater chemistry and element budgets in the Strengbach catchment (Vosges mountains, France). Water Air Soil Pollut. 79:39–59.
- Quemener, J. 1986. Nutrient balances and need for potassium p. 41–72. In Nutrient balances and the need for potassium. 13th Int. Potash Inst. Congr., France. 21–25 Aug. 1986. Int. Potash Inst., Bern, Switzerland.
- Scholefield, D., K.C. Tyson, E.A. Garwood, A.C. Armostrong, J.M.B. Hawkins, and A.C. Stone. 1993. Nitrate leaching from grazed grassland lysimeters: Effects of fertiliser input, field drainage, age of sward and patterns of weather. J. Soil Sci. 44:601–614.
- Thelier-Huche, L., A. Farrugia, and P. Castillon. 1999. L'analyze d'herbe: Un outil pour le pilotage de la fertilisation phosphatée et potassique des prairies naturelles et temporaires. Institute de L'Ëlevage, ITCF, INRA, ACTA, Chambres d'Agriculture, and COMIFER, France.
- Timmons, D.R., E.S. Verry, R.E. Burwell, and R.F. Holt. 1977. Nutrient transport in surface runoff and interflow from an aspen-birch forest. J. Environ. Qual. 6:188–192.[Abstract/Free Full Text]
- Tsirkunov, V.V., A.M. Nikanorov, M.M. Laznik, and Z. Dongwei. 1992. Analysis of long-term and seasonal river water quality changes in Latvia. Water Res. 26:1203–1216.
- Tyson, K.C., J.M.B. Hawkins, and A.C. Stone. 1993. Final report on the AFRC-ADAS drainage experiment. Agric. and Food Res. Council, Inst. of Grassland and Environ. Res., North Wyke Res. Stn., Okehampton, UK.
- Ulen, B. 1999. Leaching and balances of phosphorus and other nutrients in lysimeters after application of organic manures or fertilizers. Soil Use Manage. 15:56–61.
- USDA Soil Conservation Service. 1975. Soil taxonomy: A basic system for soil classification for making and interpreting soil surveys. USDA-SCS, New York.
- Viessman, W., Jr., and G.L. Lewis. 1996. Introduction to hydrology. 4th ed. Harper Collins College Publ., New York.
- VSN International. 2000. GenStat 4.2. 5th ed. VSN Int., Hemel Hempstead, UK.
- Zobisch, M.A., C. Richter, B. Heiligtat, and R. Schlott. 1995. Nutrient losses from cropland in the central highlands of Kenya due to surface runoff and soil erosion. Soil Tillage Res. 33:109–116.
TOP
ABSTRACT
INTRODUCTION
