Estimating Sichuan’s Key Forage Production Potential Using GIS And Forage Simulation Modeling

The potential production of two key forage species, annual ryegrass (Lolium multiflorum L.) and whipgrass (Hemarthria compressa (L.f.) R.Br.), were estimated by two forage crop simulation models (For-PASM and DairyMod) in Sichuan Province. A multiple-harvest forage model (For-PASM) was developed and validated and compared with a pasture-based model (DairyMod) to predict seasonal and multi-year forage availability. Using ArcGIS interpolation mapped For-PASM simulation annual total yield of annual ryegrass in Sichuan Province. Accurate estimates of cool-and warm-season grassland species could lead to improved management of forage-livestock systems through proper balancing of forage production with animal numbers. Spatial forage potential yield map will help with decision making for improved grasslands. If implemented, proper matching of animal numbers with forage production would better improved forage and livestock system management. The main conclusions were as follows:(1)For-PASM (Forage Potential Production Adapation Simulation Model) has created and validated. This model had normal simulation system and parameter regulation system. Each of them had several modules, including development stage and date, dynamic changes of leaf area index, harvest date and times, photosynthesis, and daily and total annual biomass accumulation. Plant development was temperature driven, with duration of growth stages dependent on accumulated heat units (AHU). Harvest date was calculated by the AHU required for boot stage development for grasses and the bud stage for legumes. Long term Annual ryegrass and whipgrass trail yield data were used to validate For-PASM simulation of seasonal production profile and annual yield in Yaan. For-PASM validation results showed that it could be used to accurately predict the seasonal production profile for annual ryegrass and whipgrass, and RMSE=1240 kg/hm2, NRMSE=7.9% for annual ryegrass annual total yield and RMSE=3147 kg/hm2, NRMSE=11.5% for whipgrass annual total yield, respectively.(2)DairyMod, biophysical pasture simulation model, composed of numerous modules with a common underlying structure. It was introduced and used mowing patterns to simulate annual ryegrass growth rate and evaluate annual total yield. Annual ryegrass trial Data collected in Ya’an and Hong’ya experimental stations to validate Dairymod results of the growth’rate and annual total yield. DairyMod validation results showed that the predicted growth rates were consistent with observed data in Ya’an and Hong’ya (R2=0.97, D=0.94, RMSE=340 kg/hm2, NRMSE=2.3% at Ya’an and R2=0.91, D=0.58, RMSE=1132 kg/hm2, NRMSE=4.5% at Hong’ya).(3)Two Forage simulation models (For-PASM and DairyMod), were compared for their ease of use and accuracy in predicting annual yield of annual ryegrass in Yaan experiment station. For-PASM simulation annual total yield of annual ryegrass and whipgrass were mapped showing greatest yields in the southeastern section of Sichuan, with a gradient decreasing to lowest production in the northwest, adjacent to Tibet. The highest yield of annual ryegrass and whipgrass were in Southwest of Sichuan, with 21500 kg/hm2 and 65 500 kg/hm2, respectively. Future work with local experts will compare predicted maps with their expert opinion and field-based trial data.