Estimating aboveground net primary production in grasslands: a simulation approach

A simulation model which reflected the seasonal dynamics of biomass in a North America mixed prairie was employed to evaluate methods of calculating aboveground net primary production from harvest data. The methods of calculating aboveground net primary production were evaluated under conditions with and without variability in the data. A multivariate normal random number generator was utilized in the variance case. In the no variance case, methods which utilized only live biomass underestimated aboveground net primary production by approximately 42%. Methods which utilized live and recent dead (senesced) biomass produced estimates very close to the model value. Variance in the data resulted in overestimation of aboveground net primary production with the degree of overestimation related to the sampling frequency. Use of a simple simulation model which simultaneously makes use of all the data to arrive at an estimate of aboveground net primary production may prove to be superior to the statistical estimators