Hi Scott,

it might be useful to look at the data produced within the framework of CORDEX (Coordinated Regional Downscaling Experiment). In CORDEX, a number of regional climate models are run as time slice experiments forced by models of the CMIP3 and/or CMIP5 ensemble with the aim to provide climate change relevant data for every land point on Earth. The results have a spatial resolution on the order of 20km I think. The data are free to use, so it might be worth a browse.

http://wcrp-cordex.ipsl.jussieu.fr/

Cheers,

Karsten

I will recommand CORDEX too since the resolution is 0.44 by 0.44 degrees but be careful about precipitation parameter. I've done some assessment of this data over Middle East and North Africa (MENA) region and found that some models overestimate this parameter while others look like presenting a lot of noise especially those downscaled by the SMHI regional model RCA4. So you have to assess those outputs to choose the best ones.

Dear Scott, as you said, the estimates are highly uncertain, so I suggest you to continue your research only on two scenarios: RCP 2.6 and RCP 8.5 . according  to IPCC AR5 WGII, it's better to study only on this two scenarios, and I think the results could be more sensitive. I'm a MSc student at university of Tehran, and I'm trying to develop an applied downscaling model using ESGF CMIP5 data, as my MSc Thesis. so I will be glad if I could help you during my activities in future.

Best regards.

CMIP5 projections are provided PER COUNTRY by the USGS here:

http://regclim.coas.oregonstate.edu/visualization/gccv/cmip5-global-climate-change-viewer/index.html

Click on a country (on the map) to see graphs displayed.

Where it says "Temperature" change it to "Precipitation" (it's a pull down menu).

Click on "Select Dataset" to choose the RCP and time horizon of interest.

You can display results for each of 26 individual GCMs, or for the average of them.

GCMs can differ hugely between them.  If in a given country you get fair agreement among most GCMs, then perhaps you can hope to attach some confidence to those projections.  But this topic of UNCERTAINTY is not what you asked about.

Mariza

Dear sir (Kenneth M Towe), if you chose CMIP3, the base period will be pre-industrial (PI), while AR4 base period was 2080-2099 (IPCC AR4 WGI, chapter 11), and WMO climate base period is 1961-1990 or 1971-2000. Also AR5 base priode is 1986-2005 (IPCC AR5 WGI, Annex I), so at the link, the base period is 1980-2004, maybe referred to AR5 base period (20-year base period).

Dear sir (Kenneth M Towe), I'm agree with you, so I think maybe the link is a simple, userfriendly site which shows default result simplified of ESGF (CMIP5 Project), also I searched about CORDEX, which its data is as another project of ESGF, and now I'm trying to compare CORDEX and CMIP5 results. I hope to be able to report you my results.

There are large uncertainties with the model-based projection of precipitation. We ever reported the simulation results from single model and multi-models for varied scenarios to the hydrological experts of the Ministry of Water Resources ten years ago, at a consultant meeting of China water resources integrated planning. They simply did not believe, and invited us to do the prediction of precipitation for the Haihe River of North China for the coming 1-2 decades using any procedures but model simulation. We did it by applying the long-term precipitation records and statistical methods, and the result has been proved to be really good.

In the Mekong River Delta of Vietnam, it is projected:

According to the low emissions scenario (B1), by the end of 21st century, rainfall is expected to increase by about 3 to 4% compared to the baseline. Rainfall is predicted to increase from 12,8 mm/year in 2020 and 14,2 mm/year in 2030.

In the medium emissions scenario (B2), rainfall tends to increase in rainy months (by up to 25% by the end of the century) and decrease in dry months (can be from 30 to 35%). By the end of the 21st century, rainfall would increase with an increase of 5-10% compared with the baseline period. By the end of the 21st century, the heaviest daily rainfall decreases at a rate of about 20 to 30%. 

"Ren... What is "long-term"? From when to when? Annual, monthly, decadal?"

"long-term" here means at least more than 100 years. Such data are insufficient in most regions, but those available would be extremely valuable. And we also applied the dryness-and-wetness grade series of the past 500 years for eastern China. The analysis results could be found in our publications including one article published in Journal of Applied Meteorology and Climatology (JAMC).