Jump to Main Content
Sensitivity of diet choices and environmental outcomes to a selective grazing algorithm
- Zilverberg, Cody J., Angerer, Jay, Williams, Jimmy, Metz, Loretta J., Harmoney, Keith
- Ecological modelling 2018 v.390 pp. 10-22
- C3 plants, C4 plants, algorithms, animal production, autumn, crude protein, diet, digestibility, excretion, feces, food choices, grasses, grazing, nutritional adequacy, runoff, sediments, seeds, simulation models, soil, spring, stems, stocking rate, urine, vegetation, water stress, Kansas
- We modified the grazing module of the APEX simulation model to allow for selective grazing of plant species and dietary-specific excretion of urine and feces. To determine the sensitivity of these changes on nutrient flows, vegetation responses, and grazer diet, we simulated the conditions of a 20-year historic grazing treatment conducted in Kansas, United States of America. Simulated scenarios were a factorial of 2 grazer types (naïve and selective) and 3 stocking densities (high, medium, low). Naïve grazers represented model behavior before our modifications. Selective grazers selected forage species based on crude protein content, digestibility, and anti-quality factors. We found that the new selective grazing algorithm resulted in a change in grazer diet quality, nutrient excretion, and standing vegetation residue. The difference between selective and naïve grazers varied over the course of the grazing season, with greatest differences occurring early in the year. C3 grasses were generally preferred in the spring, after which time C4 grasses were preferred until autumn. Simulated diet quality was relatively insensitive to stocking density, although stocking density did impact standing vegetation residue. Urinary and fecal N excretion varied seasonally and interacted with weather-induced plant water stress. The revised model’s ability to simulate differences in vegetation residue, diet quality, and nutrient excretion provide necessary building blocks to use or improve the model for the simulation of animal production, sediment erosion, water runoff, N and P runoff, and soil C accretion. The model’s performance might be improved by tracking individual plant components (e.g. leaves, stems, seeds, mast), albeit at the cost of greater input requirements.