Development and implementation of an interactive Spatial Decision Support System for decision makers in Benin to evaluate agricultural land resources—Case study: AGROLAND

Abstract

The sustainable use of agricultural land resources with concern to food security is essential, particularly in developing countries, where yields depend primarily on the biophysical conditions. To support decision making concerning national agricultural land usage, a computer based Spatial Decision Support System (SDSS) was developed. Within this SDSS, named AGROLAND, decision makers are able to visualise and evaluate biophysical agricultural land resources based on a marginality index for agricultural land use (MI). Data derived from remote sensing like MODIS or SRTM are thereby interesting and embolden sources to derive natural constraints. MI is calculated by using a fuzzy logic determination algorithm within the interactive SDSS. In AGROLAND, possibilities of user interactions during runtime as well as advanced model-based raster analyses are implemented to generate MI. This paper explicates (i) the development of AGROLAND for Benin, and (ii) the system implementation in institutions optimizing management strategies on the national scale.

Introduction

Spatial Decision Support Systems (SDSSs) contain functionalities of Geographical Information Systems (GIS) and remote sensing (RS). Additionally, they have interfaces available to access and run scientific models, and they use expert knowledge as well as logical decision trees to generate their results (Keenan, 2006, Malczewski, 2006, Laudien and Bareth, 2007). These SDSSs can be used as comprehensive computer based decision support tools for numerous problems in environmental research and management.

The work described in this paper was carried out within the interdisciplinary research project IMPETUS (an integrated approach to the efficient management of scarce water resources in West Africa) (Speth et al., 2005). One major task in the third project phase (2006–2009) was the development, implementation and application of SDSSs for the water resource management in two selected catchments in Benin and Morocco. The SDSSs of IMPETUS are implemented in a Java/XML based software framework (Enders et al., 2007).

One problem cluster of IMPETUS (which is defined as a meta-problem which requires a multi-disciplinary analysis in order to allow for drawing conclusions with respect to possible future developments (Speth et al., 2005)) addresses the evaluation of agricultural land resources in Benin (Western Africa) based on the marginality index of agricultural land use (MI). This index was originally developed by the Potsdam Institute for Climate Impact Research (PIK) and the Max Planck Institute for Meteorology (Cassel-Gintz et al., 1997, Luedeke et al., 1999), to describe and detect natural agricultural marginal sites on a global scale (spatial resolution 0.5° × 0.5°). Agricultural marginal sites are characterised by various environmental constraints limiting agricultural land use and a high risk of land degradation caused by agricultural activities. Thus, for the evaluation of the index, several natural constraints limiting agriculture under low capital input are quantified and summed to one index (see Table 1).

Remote sensing data are taken to derive some of the constraints. MODIS (Moderate Resolution Imaging Spectroradiometer) and SRTM (Shuttle Radar Topography Mission) are taken for instance to derive temperature constraints (TEMP) and high risk of erosion due to steep slopes (SL), respectively (Roehrig, 2008). Additional to the constraints, the compensation of natural aridity by irrigation near inshore waters (IC) is taken into account as it can be implemented even with low capital input. This feature was also derived by the SRTM data. As the original spatial resolution of the MI is insufficient for applications on higher spatial scales, Roehrig and Menz (2005) and Roehrig (2008) applied the MI from the global to regional (spatial resolution of 0.05° × 0.05°) and to national scale (spatial resolution of 1 km × 1 km), respectively. Therefore, investigations have been undertaken to regionalize this approach for Benin at a spatial resolution of 1 km × 1 km using influencing factors in a higher spatial resolution and an adapted fuzzy logic based algorithm. The outcomes indicated that the index can be transferred up to the national level using input data in a higher spatial resolution and an adapted membership function (cf. Roehrig, 2008). To transfer the knowledge and results to institutions in Benin, the SDSS AGROLAND was initiated in 2005 for supporting and optimizing management strategies on the national scale. As Laudien et al. (2007) described the design of this SDSS, and Roehrig and Laudien (2009) showed the used methodological approach to generate the MI for Benin, this contribution consequently focuses on the software development and implementation of the system.