Agronomie 24 (2004) 295-304
DOI: 10.1051/agro:2004037

Wheat yield estimation using remote sensing and the STICS model in the semiarid Yaqui valley, Mexico

J.C. Rodriguez^a, B. Duchemin^{b, c}, R. Hadria^c, C. Watts^a, J. Garatuza^d, A. Chehbouni^{b, c}, S. Khabba^c, G. Boulet^{b, c}, E. Palacios^e and A. Lahrouni<sup>c

a</sup>  IMADES, Instituto del Medio Ambiente y Desarrollo Sustentable, Hermosillo, Mexico
^b  CESBIO, Centre D'Études Spatiales de la Biosphere CESBIO, Toulouse, France
^c  FSSM, Faculté des Sciences Semlalia, Marrakech, Morocco
^d  ITSON, Instituto Tecnologico de Sonora, Cd. Obregon, Mexico
^e  COLPOS, Colegio de Postgraduados, Montecillo, Mexico

(Received 17 June 2003; accepted 8 September 2003)

Abstract - During the 1999/2000 agricultural seasons, an experiment was carried out on winter wheat fields in the semiarid Yaqui Valley (Northwest Mexico). This data set was used to calibrate the evolution of the leaf area index (LAI) simulated by STICS, which was found to be in excellent agreement with estimates obtained from field reflectance measurements. After calibration, STICS was able to simulate satisfactorily the seasonal levels and trends observed in net radiation, soil moisture and evapotranspiration, but the crop temperature was overestimated by about 2.5 °C. On a larger scale, STICS was run on 16 fields with contrasting management practices. The simulations indicate that yield predictability is significantly lower for later sowing dates, consistent with observations. The seasonal variations of field and satellite data (Landsat-ETM+, Terra-MODIS and VEGETATION) NDVI were very close. However, some difficulties were noted: saturation of NDVI at high LAI values and smoothed variability on a 1-km spatial scale, as well as the need for a sound methodology for processing satellite data.

Key words: crop model / STICS / calibration / winter wheat / LAI / reflectance / NDVI / remote sensing

Corresponding author: J.C. Rodriguez jcrod@imades.

© INRA, EDP Sciences 2004