Long-term monitoring of nutrient losses from agricultural catchments in the Nordic-Baltic region - A discussion of methods, uncertainties

Monitoring of nutrient losses to surface waters at the small agricultural catchment scale has been carried out in seven Nordic and Baltic countries for the past two decades, with the aim of obtaining information on agricultural activities and their impact on surrounding waters. The monitoring methods have been harmonised and in almost all 35 catchments currently being monitored, water samples are taken flow-proportionally, water discharge is measured continuously and information on crops grown and crop management is collected yearly. Here, we evaluate and discuss uncertainties in data and their consequences for fulfilling the aim of the monitoring systems. We also suggest methods and priorities for future monitoring. Although the catchments altogether cover major variations in climate, soil texture and farm type, there is a bias towards more intensive agricultural production compared with the surrounding agricultural regions. For applicability of the results to larger agricultural areas, the representativity of the selected catchments should be stated. The determination of nutrient loads is affected by uncertainties in water discharge measurements, water sampling and chemical analyses, but these uncertainties may decrease over time due to improvements in the methods used. This would improve the possibility to detect trends in the naturally large variation in nutrient loads. For wider applications, the level of nutrient loads at the catchment stream outlet must be related to catchment characteristics such as proportion of agricultural land and contribution from other land uses and point sources. These sources are often estimated based on limited information, but should be included regularly in monitoring schemes. Also data on agricultural activities are related to uncertainties such as coverage of crop management information and nutrient content in manure and harvested products. When data are missing, standard values are used, further contributing to uncertainties. However, intensive monitoring increases awareness of sources of uncertainty in data, and hence the possibility to reduce these uncertainties. Therefore, the intensive monitoring approach should be maintained and rather extended, for example by using sensors for high resolution water quality measurements so that the variations can be identified and related to catchment activities and to processes in streams and groundwater. To further understand these processes and crop management systems, modelli