Question
Asked 21st Jul, 2014

What is the best/easiest computer software for GCM downscaling, Statistical methods (BCSD & CA) ?

I am not that much familiar with the climate models work. I am civil engineering grad student working on a project to assist the impact of climate change on water resources in a small area in North California. Two downscaling methods have been adopted for the study,  Bias correction and statistical disaggregation (BCSD) and Constructed Analogue (CA).
My question here is what is the best/easiest available softwares to generate the different scenarios from the GCM to RCM applying the above two mentioned statistical methods (BCSD & CA) ?
Thanks in advance!

Most recent answer

9th Jul, 2017
Oluwafemi Adeyeri
City University of Hong Kong
@Pankaj Pandey, I have tried using climdown but it keeps complaining that it could not convert kg/m2/s of my precipitation gcm nc file. I dont know why this is hapening.
1 Recommendation

Popular Answers (1)

8th Aug, 2014
Christopher Jack
University of Cape Town
I would just add that there is a very wide range of ESD methods and which method you use can have very large consequences.  Some methods like BCSD are fairly border line in that they draw signal directly from the GCM grid scale precipitation which is widely accepted to be one of the least skillful outputs of a GCM.  Other methods use circulation based predictors which are likely to have far higher skill.  Sorry to punt a paper I'm co-authored on by it does cover these issues quite well.  Attached.
If you use the R statistical programming language there is a nice R package to support various ESD methods:  ftp://ftp.met.no/users/rasmusb/esd_0.4-1.tar.gz
9 Recommendations

All Answers (10)

23rd Jul, 2014
Mou Leong Tan
Universiti Sains Malaysia
You can just directly download the climate data that downscaled using the BCSD approach at http://www.engr.scu.edu/~emaurer/global_data/.
 I hope it helps.. All the best in your project..
1 Recommendation
24th Jul, 2014
Juan Rivera
National Scientific and Technical Research Council
You can try using the Statistical Downscaling Model (SDSM): www.sdsm.org.uk/. You just have to register and then you can freely download the model. There is a paper of Wilby and Dawson (2013) that assess the evolution of the SDSM. You may take a look: http://onlinelibrary.wiley.com/doi/10.1002/joc.3544/abstract
Have a nice day!
2 Recommendations
30th Jul, 2014
Ali Mehran
San Jose State University
Hi,
There are several softwares available online that you can use for this purpose.
I suggest developing your own code if this is a part of you project.
As you are working with climate models and you want to adjust bias, there are many issues you want to deal with.
Just be careful while downscaling you should remove noise from the product.
3 Recommendations
1st Aug, 2014
Dheeraj Kumar
Innovante Water Solutions Pvt. Ltd.
Well I will go with Juan Rivera, if you are a newbie in downscaling .
SDSM is free and simple, even you don't have to go for bias correction, SDSM handles everything for you.
I prefer using Support Vector Machine (SVM) or Stochastic approach (distribution function).
Multi Linear Regression (MLR) also gives good results.
Be sure to choose your predictor wisely.
Best Wishes
3 Recommendations
7th Aug, 2014
Vinay Kumar
Texas A&M University - Kingsville
Use simply liner regression method : OBS = MODEL*SLOPE+INTERCEPT
for details see this paper: Improved Seasonal Precipitation Forecasts for the Asian Monsoon Using 16 Atmosphere-Ocean Coupled Models. Part I: Climatology
1 Recommendation
8th Aug, 2014
Christopher Jack
University of Cape Town
I would just add that there is a very wide range of ESD methods and which method you use can have very large consequences.  Some methods like BCSD are fairly border line in that they draw signal directly from the GCM grid scale precipitation which is widely accepted to be one of the least skillful outputs of a GCM.  Other methods use circulation based predictors which are likely to have far higher skill.  Sorry to punt a paper I'm co-authored on by it does cover these issues quite well.  Attached.
If you use the R statistical programming language there is a nice R package to support various ESD methods:  ftp://ftp.met.no/users/rasmusb/esd_0.4-1.tar.gz
9 Recommendations
19th Mar, 2016
Charles Bwalya Chisanga
The Copperbelt University
You may also try to use Long Ashton Research Station Weather Generator (LARS-WG) stochastic weather generator to downscale weather data for a single site for the current and future climate forcing. url: http://www.rothamsted.ac.uk/mas-models/larswg
1 Recommendation
23rd Jan, 2017
Pankaj Pandey
North Eastern Regional Institute of Science and Technology
you may use ClimDown R easiest and best package
3 Recommendations
24th Jan, 2017
Charles Bwalya Chisanga
The Copperbelt University
Thanks Pankaj Pandey for assistance.

Similar questions and discussions

Related Publications

Article
We present a detailed downscaling simulation methodology for generating precipitation events, conditioned on external climate covariates, on a network of meteorological stations. These events can be input to hydrological models. To simulate an event on a future day t, the method uses the K-nearest neighbor algorithm to identify close neighbors from...
Article
Full-text available
Bias correction and statistical downscaling are now regularly applied to climate simulations to make then more usable for impact models and studies. Over the last few years, various methods were developed to account for multivariate – inter-site or inter-variable – properties in addition to more usual univariate ones. Among such methods, temporal p...
Article
The Coordinated Regional Downscaling Experiment (CORDEX) is a program developed by the Task Force on Regional Climate Downscaling of World Climate Research Programme (WCRP). The Task Force's mandate is to develop a framework to evaluate regional climate downscaling techniques; foster an international coordinated effort to develop improved downscali...
Got a technical question?
Get high-quality answers from experts.