A Modelling Analysis to Determine N-risk Indicator

of 4

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
PDF
4 pages
0 downs
18 views
Share
Description
Nitrogen, Modelling
Tags
Transcript
  Proceedings of the 4 th  Australasian Dairy Science Symposium 2010 221 A modelling analysis to determine N-risk indicators I. VOGELER, R. CICHOTA, V. SNOW 1 , R. MUIRHEAD 2 , and C. DE KLEIN 2   AgResearch, Private Bag 11008, Palmerston North, New Zealand 1 AgResearch, Private Bag 11008, Lincoln Research Centre, Christchurch 8140, New Zealand 2 AgResearch, Private Bag 11008, Invermay Agricultural Centre, Mosgiel 9053, New Zealand ABSTRACT The APSIM (Agricultural Production Systems SIMulator) model was used to determine risk indicators of nitrate leaching from urine patches. The simulations used climate data from three climate stations within the Canterbury area, with average annual rainfall of 600, 800 and 1200 mm, and average annual temperatures ranging from 9.8 to 11.8 o C. Six different generic soil types were used in the simulations. Deposition of a urine patch was simulated by applying an equivalent of 250 or 750 kg N/ha, and these depositions were, in separate simulation runs, done every month. The total amount of N leached was summed over three years following the urine deposition. The indicators of risk of N leaching were determined by stepwise regression and classification and regression tree (CART) analysis. The month of urine deposition largely affects the risk of N leaching, with a higher leaching risk from January to April, and a lower risk from May to December. Also N leaching was higher in the sandy soil compared with finer textured soils. The most important predictive variable for N leaching in the sandy soil was cumulative precipitation in the 3 month following urine deposition, followed by cumulative rain in the first week after deposition, and the average nitrate amount in the top 20 cm of the soil 2 weeks prior to deposition. The above results are an initial step to identify indicators of areas and timings with high risk of N leaching. INTRODUCTION An objective within the Pastoral 21 environment programme (funded jointly by FRST, DairyNZ, Fonterra and Meat & Wool NZ) is the development of a farmer-friendly tool to manage the risk for nitrogen (N) and phosphorus (P) losses to water ways and groundwater. This tool is aimed at assisting farmers to make more informed tactical and operational management decisions to minimise N and P losses to water bodies, and thus help to deliver cleaner water. A prerequisite for the development of such a tool is the determination of early indicators of risk of contaminant losses to water. These indicators must be easily measured, forecast early enough so that management action can be taken to avoid the risk, and must be appropriate for a range of environmental conditions. The risk indicators can also be used to develop additional mitigation options and management strategies to reduce losses. In this paper we present a modelling exercise performed to identify early indicators of high risk of N leaching, including site specific factors such as soil type, average annual rainfall, average annual temperature, as well as dynamic factors, such as time of deposition and rainfall following urine deposition. MATERIALS AND METHODS APSIM modelling The APSIM model (Keating et al. , 2003), utilising SWIM (Verburg et al. , 1996) as the soil module and AgPasture (Li and Snow, 2010) as the pasture module was used for determining N risk indicators. Climate data from 3 locations (here called Lincoln, Darfield, & Alford Forest) from the Virtual Climate Station database of NIWA (Tait & Turner 2005; Cichota  et al.  2008), (www.cliflo.niwa.co.nz) were used. These cover average annual temperatures from 9.8 to 11.8 o C and average annual rainfall values of about 600, 800 and 1200 mm. Nitrogen leaching data was generated using multiple runs with a complete factorial combination of these three climate stations, six different generic soils (sand, loam, clay loam, silt clay, pumice loam, pumice sand), and two N deposition rates (250 and 750 kg N/ha), and 12 separate application dates (once every month). The simulations were run for 20 different years, nitrogen fertiliser was applied at a rate of 20 kg N/ha/month from September to June, and the pasture was irrigated and managed under a cut and carry system. This cumulative nitrate leaching over 3 years after deposition was related to site specific factors, including soil type, rainfall, average annual temperature as well as dynamic factors at various times and time periods over the year, such as nitrate   concentration in the soil, soil water content, precipitation, precipitation exceeding certain values, pasture growth rate . Statistical analysis For the statistical analysis the JMP statistical package was used (www.JMP.com). Stepwise regression analysis was used to determine the factors that explained most of the variance in N leaching, and these factors were then used in a classification and regression tree (CART) analysis with directed splits.  222   Vogeler et al.– N-risk indicators modelling  FIGURE 1:  Effect of Soil Type and N deposition amount on nitrogen leaching.                                                                              FIGURE 2:  Effect of month of urine deposition on nitrogen leaching for all 6 soils and 3 climate stations simulated (left) and variability in nitrogen leaching for a sand at a deposition amount of 750 kg N/ha and clay loam at a deposition amount of 250 kg N/ha (right). ! # $% &'% $ ! ! & ' ( ) *                                                                   ! # $% &'% $ ! ! & ' ( ) *                                                                      RESULTS APSIM model simulations Simulations runs with the three climate stations, six soil types, two applications amounts, 12 application timings, and 20 years, a total of nearly 9000 different combinations, showed that N leaching was highly variable, with between 7 and 94% of the N applied leached in the three years following urine depositions of 250 and 750 kg N/ha. As expected, more N leaching occurred under the light pumice sand and sandy soils compared to soils with higher silt and clay content (Figure 1). Apart from the soil type, the month of deposition also largely affected the amount of N leaching, with a higher leaching risk from January to April, and a lower risk from May to December (Figure 2). September and October urine depositions had the lowest leaching risk. The pattern was similar for all soils. The variation in N leaching from any month was quite large, as shown for a sandy soil and a deposition amount of 750 kg N/ha and a clay loam at a deposition amount of 250 kg N/ha (Figure 2).  Proceedings of the 4 th  Australasian Dairy Science Symposium 2010 223 FIGURE 3:  Stepwise regression analysis showing cumulative predictive accuracy for N leaching, as measured by the coefficient of determination (R 2 ), after adding additional factors. +++++++       ,            *%-  . ) '%' /  0 -% '0. ) % /  / -/ '0. (    '0  /0 #-% '0. ) % /   '0. ) % /  0 -/ '0. ) '%' 1   /  0 -% '0. ) '%' 1   /  /0 -% '0,. ) '%' 1   /  0 -% '0. )%  %/   #-% '0. *% %   - '0. )% % '%% /  /0 -/ '0 To determine the factors driving this variability a stepwise regression analysis was performed for the sandy soil with the Darfield climate and a deposition rate of 750 kg N/ha. The analyses were restricted to only the most risky deposition months (January to April). Cumulative precipitation in the 3 months following urine deposition was the most important predictive variable, followed by cumulative rain in the first week after deposition, and the average nitrate amount in the top 20 cm of the soil over 2 weeks prior to deposition (Figure 3). Based on the stepwise regression results, a CART analysis was performed to identify the factors that lead to high risk of leaching. For the sandy soil the risk of N leaching from January to April was high if the cumulative precipitation in the 3 month following urine deposition was higher than 190 mm, and the nitrate amount in top soil prior to depositions was higher than 13.5 kg N/ha. The risk of leaching is lower if the cumulative precipitation in the 3 months following urine deposition was lower than 190 mm, and the nitrate amount in top soil lower than 12.1 kg N/ha. CONCLUSION To examine the environmental conditions and variables that lead to high risk of N leaching from urine patches, APSIM simulation runs with 9000 combinations were conducted, and statistically analysed via stepwise regression analysis and CART analysis. The statistical analysis determined the relative contribution of each predictor variable in explaining N leaching. For a sandy soil and a urine deposition rate of 750 kg N/ha the most important predictive variables for N leaching were: cumulative precipitation in the 3 month following urine deposition, cumulative rain in the first week after deposition, and the average nitrate amount in the top 20 cm of the soil 2 weeks prior to deposition. The month of deposition largely affected the amount of N leached, with a higher risk from January to April, and a lower risk from May to December, and the least risk in September and October. This is likely to be linked to higher uptake of nitrogen by the pasture from urea application in early springtime. The above results are an initial step to identify indicators of areas and timings with high risk of N leaching and help to develop management strategies to reduce such losses. Further analyses are however needed to fully understand the determinants of the magnitude of N leaching. ACKNOWLEDGEMENTS This work was conducted under the P21 Environment Programme, jointly funded by FRST, DairyNZ, Fonterra and Meat & Wool New Zealand.  224   Vogeler et al.– N-risk indicators modelling   REFERENCES Cichota R, Snow VO, Tait AB (2008) A functional evaluation of the Virtual Climate Station rainfall data.  New Zealand  journal of agricultural research   51 : 317-329. Keating B, Carberry P, et al. (2003) An overview of APSIM, a model designed for farming systems simulation.  European journal of agronomy   18 : 267-288. Li FY, Snow V (2010) Nitrogen processes and nitrogen use efficiency in pasture systems under different Managment - a modelling analysis. In: Proceedings of the Workshop on Nutrient Management in a rapidly changing word.Massey University, Palmerston North, New Zealand. (Eds. LD Currie). Tait A, Turner R (2005) Generating multiyear gridded daily rainfall over New Zealand.  Journal of applied meteorology   44 : 1315-1323. Verburg K, Ross PJ, Bristow KL (1996) SWIMv2.1 User Manual CSIRO Division of Soils: Canberra, Australia.
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks