Downscaling weather data

Background

Weather data are critical for many applications related to agriculture and food security.  Although computational performances have increased exponentially in the last years, there are still strong limitations in global climate modeling in simulating regional processes on short spatial and temporal time scales.  Many climate change projections are currently based on models that simulate coupled atmospheric and oceanic processes at coarse resolutions.  Resolving scale issues is a challenge to making significant progress in reducing biases in climate models. This work involves the use of a novel technique to reconstruct daily rainfall data from climatological information and to downscale the data to resolutions that are useful for local decision-making in agriculture and natural resource management.

Objectives

The objective of this work is to develop the proof of concept of a downscaling statistical model based on wavelet-multifractal (WMF) cascade approaches. Phase 1 will see the completion of time series daily rainfall at different spatial resolutions, to be used for developing downscaling parameters and building a daily weather generator.  In phase 2, the WMF approach will be applied and evaluated, and a proposal developed for global application of the approach.

Activities

Phase 1 will see daily rainfall reconstruction for sites based on NDVI and TRMM remotely sensed data, for the period 1999-2009.  Daily rainfall time series based on global NDVI coverage will then be generated, and a wavelet-based rainfall generator designed, developed and validated.  A methodological workshop is to be held in late March 2011 to discuss and compare the approach.  Phase 2 activities will include the development of a “proof of concept” for stochastic downscaling model based on wavelets-multifractals.  A comparative analysis of different downscaling techniques will be undertaken and written up, and a proposal developed for a global application of the proof of concept.

Partners 

International Potato Center (CIP); Climate Systems Analysis Group (CSAG), University of Cape Town (UCT); University of California, Santa Barbara (USA); Central Potato Research Institute (India)

Scope/Location

Method development and testing in Latin America, Africa and India

Timeframe

September 2010 – April 2011 (phase 1)

April 2011 – December 2011 (phase 2)

Project documents

Workshop on Methodological Approaches in Climate Downscaling and Weather Data Reconstruction (21-22 March 2011).
Executive summary of presentations and discussions  (PDF)