Soil organic carbon changes in dynamic land use decision models

Soil organic carbon can be sequestered through various land management options depending on the soil organic carbon status at the beginning of a management period. This initial soil organic carbon status results from a certain soil management history in a certain soil climate regime. Similarly, the prediction of future carbon storage depends on the time path of future soil management. Unfortunately, the number of possible management trajectories reaches non-computable levels so fast that explicit representations of future management trajectories are infeasible for most existing empirical land use decision models. Not surprisingly, the impact of different management trajectories has been ignored.This article proposes a computationally feasible mathematical programming method for integration of soil state dependent sequestration rates in dynamic land use decision optimization models. The possible soil organic carbon range is divided in several adjacent carbon states. For each soil management practice, location, and initial soil organic carbon state, transition probabilities of moving into a different or remaining in the same state are computed. Subsequently, these probabilities are used in dynamic equations to update the soil organic carbon level before and after each period. To illustrate the impacts of this Markov chain based method, a case study portraying Australian wheat farmers is conducted.