Spatial downscaling of precipitation from GCMs scenarios via random cascades for the Toce watershed, Italy
Abstract
We present a Stochastic Space-Time Random Cascade approach to downscale precipitation from a General Circulation Model (henceafter referred to as GCM). The study area is the Toce river basin of about 1800 sqKm in area, tributary to Lake Maggiore in Italy. Because the snowfed Toce river displays complex physiography and high environmental gradient, statistical downscaling methods are required for climate change assessment, according to the Intergovernmental Panel on Climate Change (IPCC). The Stochastic Space-Time Random Cascade model is locally tuned to downscale daily precipitation from NCAR Parallel Climate Model retrieved from the IPCC's data base. For the purpose, a 10 years series of observed daily precipitation data in 24 gaging location is used as ground truth reference. Model estimation is perfeormed using the Scale Recursive Estimation approach coupled with an explicit Expectation Maximization algorithm. Seasonal parameters of the multiplicative cascade are accommodated by statistical distributions conditioned upon climatic forcing, based on regression analysis. The major advantage of the Stochastic Space-Time Random Cascade approach deals with reproducing spatial clustering, intermittency, self-similarity of precipitation fields and their spatial correlation structure, with low computational burden. The approach is used in the Toce watershed in order to investigate water resources sensitivity to climate change.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2011
- Bibcode:
- 2011AGUFM.H21D1130R
- Keywords:
-
- 1626 GLOBAL CHANGE / Global climate models;
- 1839 HYDROLOGY / Hydrologic scaling;
- 1854 HYDROLOGY / Precipitation;
- 1869 HYDROLOGY / Stochastic hydrology