An important aim of the OPTAIN project is to derive missing information on necessary model input variables in a harmonized way to allow for a sound cross-case study assessment of NSWRM effectiveness. Therefore, in this report we provide approaches 
applicable for all OPTAIN case studies (CS) to fill data gaps. 

The specific objective of OPTAINs task 3.3 was to provide methods to cover missing input data that is required for the environmental modelling and socio-economic analysis. The deliverable includes guidelines with detailed explanations about the derivation of missing data for the CS leaders.

Based on the information provided by CS leaders in the OPTAIN milestone “MS7 Data inventory of input data for integrated modelling collected from all case studies”, the following information had to be covered by approaches provided by WP3 to fulfil the 
input requirements of the models and analysis: 1) soil phosphorus content, 2) effective bulk density, 3) moist soil albedo of the top layer, 4) USLE soil erodibility (K) factor, 5) available water capacity, 6) saturated hydraulic conductivity, 7) time series crop data.

The mapping of soil phosphorus content is based on the LUCAS topsoil dataset. During the mapping the geometric mean phosphorus content by land use types – characteristic for the region of the CS – is applied. Further required data are the LUCAS Land Use / Cover Area Frame Survey, European agro-climate zone map and the land use or land cover map of the CS – a local one, if available.

For the calculation of soil physical and hydraulic properties we apply methods available from the literature.

The derivation of crop maps is based on remote sensing data. A crop classification model was trained on the cropland data of the LUCAS Land Use / Cover Area Frame Survey of the years 2015 and 2018, merged with the Sentinel-1A and -1B satellite radar images. The pixel based crop classification was carried out with a random forest algorithm on the Google Earth Engine platform. The method can be applied for 2015 and all following years. By adding a map of field boundaries, the pixel based crop prediction can be 
aggregated to field level using the majority of the predicted crop.

Regarding the socio-economic data, missing information is planned to be covered from official statistics. The EU database does not account properly for the Norwegian and Swiss sites, therefore required data will be retrieved ex novo from local sources or literature.