Geosciences Colloquium: Development of precise precipitation data APHRODITE and climate superensemble

Zoom: https://tau-ac-il.zoom.us/j/83294569872?pwd=WmdPUWRPdGVQejIvcXhsQmFJY094UT09

​Abstract:

I introduce the rain-gauge-based grid precipitation data APHRODITE and show results of applying the synthetic super-ensemble (SSE) method to seasonal precipitation over Asia and the Middle East. 

1. Asian Precipitation – Highly Resolved Observational Data Integration Towards Evaluation of water resources (APHRODITE) 

2. Multi-model Super-ensemble (MSE) by Prof. Krishnamurti (FSU)

3. Synthetic (climate) Super-ensemble (MSSE) by FSU 

4. Issues of applying Super-ensemble for temporal variations and for warming climate

The APHRODITE precipitation data is widely used for understanding monsoon variability, driving hydrological models, downscaling, and validating precipitation estimates. Since the rain-gauge products are more accurate than those of satellites, they are used as “teacher” for the MSE. It is a unique method to combine several model outputs and precise observation data to make the best forecast. Super-ensemble technique can be applied for CMIP-type climate data by adopting a singular-value-decomposition (SVD) analysis. This synthetic super-ensemble (MSSE) has successfully improved seasonal precipitation forecast over monsoon Asia.

We show the application of MSSE to the Middle East precipitation. We used the simulated precipitation of the five coupled general circulation model (CGCM) outputs, which are part of the CMIP5 project. We used the models and the APHRODITE precipitation for the 10 years of (1997/1998-2006/2007) winter months (December, January and February). For the seasonal climate forecasts, a MSSE technique and a cross-validation technique were adopted, in which the year to be forecasted was excluded from the calculations for obtaining the regression coefficients. As a result, seasonal precipitation forecast was considerably improved by the use of APHRODITE data. These forecasts are much superior to those from the best model of our suite and ensemble mean. However, unfortunately, MSSE does not represent the large-scale decreasing trend pattern, except in the eastern part of Turkey and a part of Israel.

We applied MSSE for spring precipitation over the Central Asia. To overcome the weakness of temporal analysis, we determined weighting coefficients using two sets of models: a) the six models best able to represent horizontal variation, and b) the six best able to represent temporal variation, over an 11-year period (1980–1990) with AMIP-type historical data from CMIP5. Using the different weighting coefficients, we applied the MMSE to the RCP 8.5 scenario, as it pertains to the period 2100–2109 over CA. Model set a) showed a trend toward increasing precipitation in some spring months, while set b) showed a robust decreasing precipitation trend in all months. On the basis that it is reasonable to place greater trust in model outputs that represent the variation in interannual precipitation in CA well, we predict that spring precipitation in the 2100s will be reduced in the mountain region of CA.
 

Event Organizer: Dr. Roy Barkan