Detection of bare soil on arable land with high temporal resolution as input to water quality modelling




With water management in view, the Swedish sub projects within the GMES projects AquaSAGE and GSE Land (financed by ESA) and “FP6 IP geoland” ( ) have focused on the integration of remote sensing derived information on land cover and land use into water quality modelling. To these formal GMES initiatives can be added projects of local, regional or national character. The Swedish National Space Board has in the past co-financed several of them, for example RESAP, WFD-REMGIS and Phosphorus Load Modelling Based on MERIS Data.

The need for improved information on the spatial distribution of bare soil during winter was identified in the project on Phosphorus Load, in the Rönneå drainage basin in Southern Sweden. Data from the MERIS sensor (ENVISAT satellite) with 300 m spatial resolution was used. One of the conclusions was that MERIS data has too poor a resolution for Swedish conditions (relatively small fields).

In FP6 IP geoland, the spatial and temporal resolution was improved using data from the new AWiFS sensor (IRS-P6 satellite). The sensor has a spatial resolution of approximately 50 m, the return period is short and the area coverage per registration large. One purpose was to evaluate the use of AWiFS data for operational purposes. The results are promising, data proved to be of good geometric quality and data delivery reliable. In geoland, Metria developed a technique using AWiFS data for detection of fields where tilling had occurred between two registration dates. A small test site north of Uppsala was used and the data was at that stage not integrated into a hydrological model.

The Swedish Meteorological and Hydrological Institute (SMHI) is presently developing a New Hydrologic Model Generation – NHMG – with a semi-distributed approach. The new model is developed with focus on application flexibility and scenario modelling, e.g. reduction of transport of nutrients to rivers, lakes and the sea. Within this development work, integration of up-to-date, remote sensing based information on land cover and land use is foreseen.

The water quality modellers have expressed a high interest in actual time series of tilling practices, i.e. pinpointing within a narrow time span if and when during spring or autumn the different fields in a drainage basin are planted, harvested and ploughed; this is a critical factor in determining especially phosphorus leakage from field erosion.

Project idea and goals

The project idea is to make operational the specific methodology on detection of bare soil on arable land developed within the FP6 IP geoland, apply it on several occasions during spring and autumn on a large scale and to incorporate the results in the NHMG. The model will be adapted accordingly, i.e. to the new type of data with higher spatial and temporal resolution than presently used.

The river Vindån drainage basin (~ 300 km2), south of Valdemarsvik, will be the study area. The basin is used for the NHMG development and SMHI has a dense monitoring system (in-situ) running. The co-operation with end users is also well established through the DEMO and Kaggebo projects (

With an operational modelling approach in view, that integrates results from stable and repeatable remote sensing techniques and a semi-distributed nutrient leakage model, the project goals are to:

• To consolidate the methodology for bare soil detection developed within the FP6 IP geoland and adapt it to the NHMG (temporal adaptation)

• To implement the classification methodology in the test site Vindån, Counties of Östergötland and Kalmar, with temporal adaptations

• To adapt the NHMG to the data with improved spatial and temporal resolution provided by the remote sensing approach

• To validate the model results and evaluate the new approach compared to present “state-of-the-art”

• To make cost-estimates for up-scaling to Southern & Central Sweden.
Metria Miljöanalys will carry out the remote sensing part of the project and SMHI the hydrological model development and validation.

Kerstin Nordström, Metria Miljöanalys,

Lotta Andersson, SMHI,


Senast uppdaterad: 25 maj 2009