Insufficient geographic characterization and analysis in the planning, execution and dissemination of agronomic research?

Abstract

Understanding spatial variation in crop response to environment and management is an essential component of agronomic research. Given the increasing availability of geographic information systems (GISs) and spatial data, one might anticipate widespread use of maps, reference to geographic variation, and analyses that capitalize on the power of GIS. Such use would be reflected in improved selection of research sites or treatments, better understanding of the influence of climatic or edaphic factors on crop responses, and more easily interpretable, quantitative results presented through maps. With the notable exception of research on individual fields or local landscapes, however, published research indicated insufficient use of spatial information and analyses or even explicit consideration of the geographical context of research. To assess use of geographic information in agronomic research, we examined papers in five prominent agricultural journals for evidence of analyses at a spatial scale smaller than field plots or local landscapes but larger than national or continental levels, which we term “mesoscale”. Use of simulation models was also examined since models can quantify response to environmental factors and thus, their use might provide instructive comparisons with use of spatial analyses. Of 250 papers considered, less than half (119 papers) described the geographic context of the research, and only 90 gave geographic coordinates with sufficient precision to locate sites within a 10 km radius. Only six papers included maps of the study area. Over 150 papers used single locations, and just 26 papers, more than four sites. In experiments creating variation in soil conditions, such as through irrigation, tillage, or nutrient treatments, the limited number of treatment combinations appeared to constrain quantitative interpretation of results. Most papers involving simulation models focused on model development or validation; models were seldom used to analyze effects of climatic or edaphic conditions as a complement to other lines of research. We conclude that there is need to increase the geographic relevance of agronomic research. Suggested steps include selecting sites using an explicit sampling strategy, using larger numbers of sites, and analyzing and presenting results using tools such as GIS and crop or ecosystem simulation models.