You can find our Matlab code in the answers of this question.
Let me know if you need help.
https://www.researchgate.net/post/In_bias_correction_Quantile_mapping_is_not_improving_my_results2
You can find our Matlab code in the answers of this question.
Let me know if you need help.
https://www.researchgate.net/post/In_bias_correction_Quantile_mapping_is_not_improving_my_results2
Dear Luca!, your code is good for precipitation, however, can you make a review for the temperature ? Thank you
Dear Luca, if you not mind can you share the code for us about temperature? I would appreciated for your help. Since I am not good in programming. Thank you very much ! My email: [email protected]
Try this one: https://www.researchgate.net/publication/271703152_Cumulative_Density_Function_CDF_matching_approach?ev=prf_pub
Data Cumulative Density Function (CDF) matching approach
Dear Luca !
Your currently codes QMAPP and DBS have the same results, when i compare temperature during the decades of 2020s, 2050s and 2080s to the basedline period, it does not changes too much. For monthly temperature, it change between -1C to 1C only, and for annual temperature it does not change (0C) for the three decades. Mostly temperature in my study area should be changes from 0.5-3C in the future. This makes me fell uncertainty in my results. As well as i known that from DBS code, rainfall was used 2 gamma fit distribution while, temperature used normal distribution (mean, standard deviation). However, from the code it only used for rainfall and excluded for temperature.
It depends on how you are applying the codes. You shouldn't get the same results...
I guess that you are applying the cdf matching also to future data wrt present data. It is wrong. You should apply the same transformation you obtained for present data (by matching observed and present time gcm data) to future projections.
Thanks Luca for very useful comment.
I used the RCM of HadGEM3-AO model(1956-2005 for historical data, and 2015-2100 for RCPs).
For calibration, I used observed meteorological data (rainfall/temp) (from 1977-2001) and applying with the historical data of RCM (from 1977-2001). For projections, I used again observed meteorological data (1977-2001) and applying with 25 years of RCPs in the future (2020s,2050s and 2080s). I dont know that did I do a wrong ways or not, could you please give me a suggestion?
Thank you very much Luca!
The problem might be the code, that actually it doesn't allow to extrapolate the "matching/rescaling function" from present to future data. You have to calculate the rescaling function from present data (obs and gcm), and then apply this function to future data. You do not have to rescale future data with present data.
Hope it helps...
Dear Amir,
You may find this page useful.
https://meteo.unican.es/trac/MLToolbox/wiki/Downscaling/BiasCorrection/Methods
Recently, an excellent book came out on Statistical Downscaling which you may find useful:
Book Statistical Downscaling for Hydrological and Environmental A...
Thank you
- Badges
- Science topic