SARRA-H Crop Model______________figfigfig______________Cirad, UMR TETIS, C.Baron , 2013_fig
Système d'Analyse Régionale des Risques Agroclimatologiques Version H (SARRA-H;System of Agroclimatological Regional Risk Analysis)

Translation: Peter Biggins ............................................................................................................................................................

Sarrah Historique Le modèle Téléchargements Partenariats et liens Publications Projets
Home History The Model Download Partnership Publications Projects

Fig Background

Today

SARRA-H is a major development of the SARRA software suite (see below) which was a simple dynamic water balance used to estimate the impact of a climate scenario on an annual crop. Like SARRA, SARRA-H is more specifically adapted to analysing how the growth of dry cereals and their potential yields are affected by the climate in a tropical environment (Dingkuhn et al. 2003; Baron et al. 2005; Sultan et al. 2005). The crop model simulates potential yield under water constraints by incorporating the processes of soil water balance, evaporation, potential and actual transpiration, phenology, potential assimilation and, under water constraints, the processes of maintenance respiration and, lastly, biomass distribution (leaves, stems, roots, grains). This model has performed well in climate impact analyses for tropical cereals (Mishra et al. 2008; Oettli et al. 2011). It has been calibrated with a series of local and modern varieties (of millet, sorghum and maize) through trials in controlled environments. Multi-year on-farm agronomic monitoring has been set up at several sites contrasting as much through their agricultural practices as through their climate (Niger, Senegal, Mali, Burkina Faso), thereby enabling an assessment of the predictive capacity of the model in smallholder environments (Traoré et al. 2010). Through these trials and monitoring operations, it has been possible to characterize different varieties and practices and define parameters linked to the model, thereby offering a range of scenarios representative of agricultural practices. In particular, on farm monitoring has confirmed the predominance of local sorghum and millet cultivars, which stand out through their high sensitivity to the photoperiod, which has been the subject of specific studies resulting in a module integrated into the model (Kouressy et al. 2008, Dingkuhn et al. 2008).

SARRA-H manages a library of modules that integrate the conventional principle of water constraint (water balance) and combines it with potential growth, which depends on solar radiation and its interception by the plant cover (carbon balance). It is a multiplicative model (water resource x radiative resource), completed by a phenological module (phenology) to structure the vegetation cycle and the processes linked to each phenological stage, through a simple but dynamic description of the cover (a “big leaf” characterized by morphological and geometric coefficients), and through physiological yield build-up (source-sink competition). The extrapolation domains for this model therefore cover the situations limited by water and/or solar radiation, taking into account an overall soil fertility status that is constant throughout the cycle of the plant.

Over the years

At the beginning of the 1970s, P. Franquin (1971) laid the foundations of rainfall frequency analysis. Franquin and Forest (1978) developed a water balance model whose outputs could be considered in frequency terms. Lastly, encouraged by F. Forest, several researchers from CIRAD took another look at these models and methods and enriched them from a theoretical and ergonomic viewpoint. Worth mentioning are J.C. Legoupil, F.N. Reyniers, J. Imbernon, J.P. Fréteaud, B. Lidon and S. Sabadie, who were involved in developing an initial coherent family of software programs in FORTRAN.

All these tools were widely disseminated and have served as the basis for a great deal of research and development work in the Tropics, in both Africa and Latin America. On a plot scale, the utilization principles have concerned agronomic diagnosis and irrigation management. On a regional scale, agricultural potential zoning, sowing date optimization studies and yield forecasting programmes have been implemented in numerous countries.

The advances made in computer hardware and languages have enabled the development of software programs adapted to increasingly specialized and demanding requirements. For instance, between 1987 and 1992, F. Forest and F.N. Reyniers oversaw the creation of some new tools in the Water Management Research Unit:

fig

This work was continued by other researchers, notably by F. Affholder, S. Marlet, B. Muller, P. Perez, B. Rapidel, E. Scopel and M. Vaksman, who helped to diversify the approaches and calculation algorithms. All this work was synthesized and documented by C. Baron, A. Clopes and F. Maraux in a suite of SARRA software programs, ranging from climate analysis to regional zoning, applied to the analysis of agro-climatic risks. That software suite was translated into several versions: firstly Portuguese and Indonesian, then Spanish and English.

In practice, these software programs proved to be very efficient, with the parameterization choices being a wise compromise between extreme simplicity (leading to over-coarse representations of phenomena) and excessive complexity (leading to problems in assembling the sets of parameters needed for simulations). In Brazil, in partnership with EMBRAPA, SARRA has been used since 1996, as part of the ProAgro programme for annual monitoring of agro-climatic risks for optimum sowing zoning, launched by E. Assad and followed up by F. Massena from EMBRAPA. Each year, that programme, monitored by the Ministry of Agriculture (Ministério do Desenvolvimento Agrário – MDA), brings into play 24 EMBRAPA institutes and research centres from the different States in Brazil.

fig

The new version, SARRA-H, launched by M. Dingkuhn in 2000, incorporates a dual approach, water efficiency and solar radiation efficiency, to simulate biomass dynamics based on more physiological concepts and processes. This new version still follows a simple model that is not very demanding in input data and parameters. Several researchers from CIRAD were involved in designing this new version, including M. Dingkuhn, C. Baron, B. Muller, M. Vaksmann, JC Combres, etc. At the same time, C. Baron launched a model development and simulation platform integrating a series of tools and interfaces (ECOTROP): graphics, queries, data imports/exports, database, sensitivity analysis, etc. This development was made possible through a major contribution from P. Reitz (from LIRMM) and V. Bonnal, G. Aguilar and J.C. Soulie (CIRAD) and some trainees. The different versions of SARRA H have notably been incorporated into this platform.

fig

*Diagram of the different processes simulated around the water balance *

SARRA-H and its environment have clearly evolved during projects in which the model and its use/ergonomics were tested in a large number of situations by users from varied fields of research and applications. It has thus been possible to calibrate the parameters for several selected and local (farmers’) varieties of millet, sorghum and maize. The predictive capacity of the program has been checked in a large number of situations (various trials under controlled conditions and on-farm multi-year monitoring). And of course, all this work has contributed to improving the simulation modules and processes using the results of these studies geared notably towards the impact of climate variability and change, sowing practices and seed losses, irrigation protocols, simulation of consecutive seasons, etc. Among the partners worth mentioning are S. Traore, A. Alhassane, H. Songoti (AGRHYMET-Niger), M. Kouressy (IER-Mali), B. Sarr (CERAAS-Senegal), L. Some (INERA-Burkina Faso), Benjamin Sultan (IRD), etc. and, of course, some trainees and especially PhD students (who can be found in the publications).

fig

Result display interface: calibration of a Millet trial on the dynamics of aboveground biomass, leaves and yield (observed, dots; simulated, lines)

Home History The Model Download Partnership Publications Projects
Sarrah Historique Le modèle Téléchargements Partenariats et liens Publications Projets