Journal Article
Geoscientific Model Development, vol. 15, iss. 2, pp. 429-447, 2022
Authors
Katherine V. Calvin, Abigail Snyder, Xin Zhao, Marshall Wise
Abstract
Abstract. Future changes in land use and cover have important
implications for agriculture, energy, water use, and climate. Estimates of
future land use and land cover differ significantly across economic models
as a result of differences in drivers, model structure, and model
parameters; however, these models often rely on heuristics to determine
model parameters. In this study, we demonstrate a more systematic and
empirically based approach to estimating a few key parameters for an
economic model of land use and land cover change, gcamland. Specifically, we
generate a large set of model parameter perturbations for the selected
parameters and run gcamland simulations with these parameter sets over the
historical period in the United States to quantify land use and land cover,
determine how well the model reproduces observations, and identify parameter
combinations that best replicate observations, assuming other model
parameters are fixed. We also test alternate methods for forming
expectations about uncertain crop yields and prices, including adaptive,
perfect, linear, and hybrid approaches. In particular, we estimate
parameters for six parameters used in the formation of expectations and
three of seven logit exponents for the USA only. We find that an adaptive
expectation approach minimizes the error between simulated outputs and
observations, with parameters that suggest that for most crops, landowners
put a significant weight on previous information. Interestingly, for corn,
where ethanol policies have led to a rapid growth in demand, the resulting
parameters show that a larger weight is placed on more recent information.
We examine the change in model parameters as the metric of model error
changes, finding that the measure of model fitness affects the choice of
parameter sets. Finally, we discuss how the methodology and results used in
this study could be used for other regions or economic models to improve
projections of future land use and land cover change.
