Spatializing crop models for sustainable agriculture

Abstract

Crop models mathematically represent dynamic point-scale interactions between plant, weather, soil and management practices. They have been increasingly applied large scale (i.e. from farm-level to regional and global applications) to understand and quantify the trade-off between productivity, management and the sustainability of cropping systems, in terms of responsible use of resources (e.g. water and nitrogen) and of adaptation to or mitigation of climate change impacts. This contribution reviews the most recent information about spatializing crop models and provides a comprehensive overview of major assumptions and criticalities related to this methodological approach. The first paragraph focuses on the definition of crop models, presenting their historical evolution and main fields of application. A bibliometric analysis was carried out on 1017 scientific papers published between 1990 and 2018 in order to identify the most frequent scientific topics concerning the adoption of crop simulation modelling for sustainable agriculture. The second section describes the main sources of uncertainty in spatializing crop models, addressing two main aspects. Firstly, basic assumptions and validity domains of processes/phenomena represented may still not be valid when applied in a different spatial resolution. Secondly, reference input data needed to characterize the cropping system under study, to run models and test their performance at large scale can often be scarce and/or uncertain due to aggregation/disaggregation issues. The third section defines the minimum amount of data about environment (i.e. site, weather, soil), management (e.g. sowing and harvest date, cultivars and crop operations adopted) and crop type, needed to operate crop models at a given location under current/future climate scenarios. Necessary methodological indications for building a multi-layer georeferenced database facilitating coupling with biophysical models are also provided. Ways of integrating proxy variables (e.g. obtained from pedo-transfer functions and remote sensing data) and crop models have been reported. The last section presents two case studies dealing with the spatialized application of crop models to promote the sustainability of agriculture. A European case study is centred on the definition of farmer adaptation strategies to alleviate climate change impacts, while a regional case study evaluates the efficiency of water management and water footprint of tomato cultivation in Southern Italy.