Cloud Feedbacks



Other MIPs



Google Calendar

Lab Calendar

Site Map


Privacy & Legal Notice

Thanks to Our Sponsors:

PCMDI > WCRP CMIP3 Model Output > Diagnostic Subprojects Printer Friendly Version
<< Back to WCRP CMIP3 Subprojects

Feasibility of downscaling Maritime Continent precipitation and circulation under present and future climate conditions

PI: Dan Vimont
Institution: University of Wisconson
We will use empirical downscaling techniques to debias and downscale precipitation over Indonesia and the Maritime Continent under present and future climate forcing. A characteristic of coupled climate model simulations is the strong bias in regional precipitation and tropical hydrology, despite the increasing fidelity with which they represent the large scale circulation. These biases occur despite robust observed relationships between the large scale circulation and regional precipitation (e.g. the monsoon circulation, or relationships with ENSO variability). We intend to investigate the feasibility of using empirical downscaling models (EDMs; e.g. Widmann et al., 2003) to
debias and downscale regional precipitation over Indonesia and the Maritime Continent. Preliminary analysis shows that this approach is very useful for debiasing the annual cycle of precipitation over Indonesia in coupled model simulations.

The EDMs will be constructed using observations of the large scale circulation (from the NCEP reanalysis; Kalnay et al., 1996) and regional precipitation (from station data, and from the U. Delaware Climatologically Aided Interpolation data set; Willmott and Matsuura, 2001). We will use large scale fields which are known to have physical relationships with precipitation (e.g. upper and lower level zonal winds and the monsoon shear line; SLP and the Southern Oscillation; and lower level specific humidity and the hydrological cycle). These EDMs
will be applied to model output from the present-day IPCC AR4 simulations, as well as future simulations (we will focus on the A2 and B1 scenarios). One problem with this approach is that we cannot test how well EDM that are developed using present day analyses represent future precipitation in the observed record. Thus, we will also develop EDMs "within the model world", using the same large-scale predictors from each present-day model simulation, and precipitation from the same model simulations. These EDMs will then be used to "downscale" the future precipitation to determine which large-scale
variables are most important for training the EDMs.

    Add Publication

    << Back to WCRP CMIP3 Subprojects
    For questions or comments regarding this website, please contact the Webmaster.
    Lawrence Livermore National Laboratory  |  Physical & Life Sciences Directorate