Geospatial modelling to assist the environmental management of sustainable aquaculture across large catchments.

Lynne Falconer: Ph.D.

The aim of this PhD project is to use Geographical Information Systems (GIS) and remote sensing tools to develop models that can be used to assist the environmental management of sustainable aquaculture across large catchments. The models can be used to assess the potential for site selection, seasonal changes in land use and the potential for inputs of excess nutrients to aquatic systems from the surrounding environment. This PhD project is partly funded by the EU framework 7 SEAT (Sustaining Ethical Aquaculture Trade) project and is included in Work Package 4 (Environmental Models). The SEAT project aims to explore the sustainability of aquaculture products from Bangladesh, China, Thailand and Vietnam with a focus on tilapia, Pangasius catfish, shrimp and prawns. Figure 1 shows the study areas which have been defined within a geographical context using watersheds.

Figure 1 : Study areas in Asia.

Land is often a source of conflict with different stakeholders competing over use, space and resources. Since aquaculture is intrinsically linked with the environment it is essential that the surrounding land use of the wider catchment is considered within environmental management plans. Seasonal land use models were developed for each area using a Bayesian classifier to classify Landsat ETM+ satellite imagery (Figure 2). The results of the models have shown that there are significant differences in land use and resource availability between seasons, highlighting the need to evaluate land use on a seasonal rather than annual basis.

Figure2: Figure 2: Seasonal land use models for each study area.

The variability of land use between seasons also influences the "non-point source pollution" generated by different categories of land use. Unlike "point source pollution" (single, identifiable sources such as pipes), "non-point source pollution" is difficult to identify and measure as it is generated from diffuse sources with no single point of entry and can be widespread across a large area. "Pollutants" include nutrients, sediments, oils, salt, metals and toxic chemicals. The input of excess nutrients, such as nitrogen and phosphorus, can have a significant impact on water quality and therefore could have potential negative consequences for aquaculture. Spatial models have been developed to assess the risk of "non-point source pollution" in the form of nitrogen and phosphorus across each of the four study areas for both the dry and rainy seasons (Figure 3). The models use data on land use, soil, rainfall, topography and hydrological conditions which are then combined within a weighted multi-criteria evaluation to produce the final outputs.

Figure 3: "Non-point source pollution" models.

These "non-point source pollution" models are a useful decision making tool as they allow stakeholders to identify areas where there is a higher/lower risk of potential nutrient enrichment. This provides extra information for site selection studies as planners can identify areas where there is the potential for poor water quality/eutrophication which would be unsuitable for a new aquaculture development. The models can also be used in environmental management plans to identify areas where regular monitoring should take place. Another advantage of the models is that they can be easily updated and adapted for different scenarios, such as increased rainfall and changes in land use, or updated for other "pollutants" such as sediment (Figure 4).

Figure 4: Examples of how the model can be adapted for different scenarios and other "pollutants".

Acknowledgements: This PhD is funded by the Institute of Aquaculture, University of Stirling ( and the EU framework 7 SEAT project.