Stochastic Arable-Nitrogen Abatement from Heterogeneous Sources Cost-E ective Coastal Nitrogen Abatement

The implications of the propagation of uncertainty for the e ciency of stochastic water pollution control were analyzed. The empirical analysis focused on nitrogen pollution of the Baltic Sea from Swedish crop farms. A two moment model (mean and variance) of the distribution of coastal load was developed and integrated with an economic model of arable land-use choice. It was shown that uncertainty propagates through the system if arable emissions and retention are negatively correlated, and riverine loads (emissions less retention) positively correlated. Under these conditions a precision requirement increases the e ciency of abatement in catchments with relatively high load variability ceteris paribus. Relative abatement increased signi cantly with the precision requirement in high retention regions because these regions had a relatively large impact on the variability of coastal load. Thus if the Baltic damage function is convex then increasing abatement in high retention regions should increase overall economic e ciency.