POUTEAU R. 2008. Frost risks and climate scenarios in the southern Bolivian altiplano: spatialization of quinoa crops ecological vulnerability
Article edited: Wednesday 2 July 2008 - Latest addition: Friday 2 July 2010
All the versions of this article: [English] [Español] [Français]
Study realized by Robin POUTEAU under the supervision of Serge RAMBAL and Thierry WINKEL (CNRS/IRD, Montpellier, France) using satellite data to: i) spatialize over the entire Salar of Uyuni region frost risks at a local scale taking into account the effects of topography, ii) consider the evolution of these risks by 2100 using a global scenario of climate warming.
Abstratc: The success of quinoa in the international market leads the farmers of the Salar region (southern Bolivian highlands) to cultivate in flat areas with important night radiative frost risks. Though quinoa is adapted to the extreme conditions of the Andean highlands, its frost sensitivity depends on its phenology and frost risk is never nil. The three major quinoa phenological phases (vegetative growth, flowering and maturing) are associated with contrasted seasonal atmospheric dynamics controlling frost risk. Though varying with time, these risks also depend highly on topoclimate. MODIS satellite remotely senses surface radiative temperature and cloud cover daily and everywhere in the study area with a 1 km resolution. Combined with a geographical information system and with 3 meteorological station datasets, MODIS was validated and used to map frost risks, based on a statistical approach. For each of the three quinoa phenological phases, a simple estimation model of frost risks was built from their relation with orotopographical descriptors in order to understand the processes controlling frost events and to downscale the resolution until 100 m, which is closer to the scales of the physical processes themselves and of the farmer crop practices too. In this mountain area (3650-5900 m) local climate depends highly on elevation but also on latitude, salar distance, slope and aspect. A compound topographical index was also added to quantify the adiabatic drainage. Models were calibrated by multiple regressions with a sample of topographically homogeneous pixels, before being statistically validated. This approach permitted to highlight the relative effect of each orotopographical factor on climate and their interrelations. Finally, the CNRM-CM3 global circulation model, simulating the IPCC-SRES A1B climatic scenario, was used to show that, from 1980 to 2100, the Andean highlands will be particularly affected by the global warming with a 3 °C average increase during all the crop season. This warming will result in a considerable reduction, and even total disappearance, of frost risks by 2100.
Key words: downscaling; landscape; remote sensing; radiative surface temperature; topoclimate
