Cost-effective abatement of non-point source nitrogen emissions - The effects of uncertainty in retention.

Non-point nitrogen discharges from agriculture are difficult to regulate, because of the diffuse nature of the pollution. Inflexible and uniform regulation policies have been the solution in many parts of the world. A more targeted and flexible regulation, adjusted to the heterogeneity of hydrological conditions and ambient water quality, as well as the heterogeneity in abatement costs between farms, has been challenging to develop and implement. One reason is the difficulties measuring the hydrological conditions on a detailed spatial scale. One of the most important hydrological factors, which co-determines the load reduction, is the retention (attenuation) of nitrogen in the catchment, from the root zone to the coast. It is therefore critical to understand how uncertainty in retention affects policy recommendations. In this paper, we use a spatial costs-minimization catchment model, TargetEconN, applied to the Danish catchment Limfjorden. The model includes retention, effects of measures and abatement costs at a detailed spatial scale. The model identifies optimal spatial allocation of nitrogen abatement measures at different load reduction targets to the fjord. We use the model to evaluate the sensitivity of the results to different forms of uncertainty in retention by running scenarios. For all scenarios, the total costs, the marginal costs and the distribution of measures are compared, and special emphasis is paid to how uncertainty on the retention affects the cost-effective allocation of abatement measures. The results indicate that taking spatial heterogeneity of retention into account is important to obtain cost-effective nitrogen abatement as this reduces the costs by approximately 25 percent. We assess the importance of uncertainty in the retention estimates by comparing baseline results with the results from model runs using 1) the mean retention estimate for all areas, 2) misspecification of the retention estimates by 10-20 percent in low and high retention areas, and 3) random misspecification of the retention across the catchment. The results indicate that this range of misspecification and uncertainty in retention, does not play a major role for the allocation of measures, nor for the total costs. We conclude that considering spatial heterogeneity in retention is important for cost-effectiveness, and that a policy where retention is included in the allocation of measures is relatively robust towards the uncertainty in the measurements of retention. Uncertainty and misspecification can lead to higher costs for individual farmers. However, for the agricultural community as a whole it is much more costly not to take differences in retention into account.