Plant parameterization and APEXgraze model calibration and validation for US land resource region H grazing lands

CONTEXT: Grazing lands account for 66.7% of all agricultural land in the United States. Soil erosion, water quality and conservation have been identified as primary resource concerns on grazing lands. Biophysical models driven by daily climatic variables enable assessment of natural resource management options over time and across large landscapes, especially where on-the-ground assessments are not feasible. OBJECTIVE: The objectives of the study were to develop plant functional groups to parameterize, calibrate and validate the APEXgraze model. The calibrated and validated model was applied to demonstrate its potential to evaluate environmental resource concerns and ecosystem services on grazing lands as impacted by increased cattle stocking densities, initially focusing on Land Resource Region (LRR) H -the Central Great Plains. METHODS: Because of the high plant species variability in grazing lands, APEXgraze was calibrated using plant functional groups. A total of 64 plant functional groups were identified as representing plant species in the Central Great Plains. When the functional group parameters were incorporated into the APEXgraze model, alongwith the soils data, daily weather, and a zero-grazing management system, the functional group parameters gave a reasonable representation of the plant species growth dynamics, to include leaf area development and biomass yields. The calibrated and validated model was applied to demonstrate its potential application to evaluate the impacts of high cattle stocking densities on evapotranspiration, surface runoff, water stress, water and wind erosion, and soil organic carbon storage. RESULTS AND CONCLUSIONS: Overall, model evaluation metrics indicated satisfactory performance by APEX -graze during calibration and validation against evapotranspiration (ET) data for the contiguous United States and National Resources Inventory (NRI) reconstructed field-measured biomass yields. Mean simulated ET values were reasonable, and within 10% of observed ET values. The model was able to explain 98% (R2) of the variance in observed ET values in both calibration and validation. In addition, Nash-Sutcliffe Efficiency (NSE) was >0.50, and Willmott's d was closer to 1 (> 0.80) in both calibration and validation. Aboveground biomass calibration and validation metrics were not as strong as those for ET. Still, overall model evaluation metrics indicated satisfactory performance by APEXgraze. Simulated mean biomass yields were respectively within 16% and 11% of measured biomass yields during calibration and validation. The model was able to explain 75% (R2) of the variance in measured biomass yields during calibration and 70% of the variance during validation. Both NSE and d were respectively >0.50, and closer to 1 (> 0.80) in both calibration and validation. APEXgraze's utility was successfully demonstrated by its potential application to evaluate the impacts of high cattle stocking densities on evapotranspiration, surface runoff, water stress, water and wind erosion, and soil organic carbon storage on grazing lands. Significance: Following successful calibration and validation, the APEXgraze model was applied to demonstrate its potential application to evaluate environmental resource concerns and ecosystem services on grazing lands as impacted by increased stocking densities.