Dear Saima, Hi;

I don't know what do you mean by the first part of your question, but, for the second part I want say, yeah, it's ok to perform bias correction for future data using 20 year observed data at least, but be care of the last year. you could easily check IPCC AR5 WGI, (for example the attached file is: Annual mean changes in precipitation (P), evaporation (E), relative humidity, E – P, runoff and soil moisture for 2081–2100 relative to 1986–2005 under the Representative Concentration Pathway RCP8.5, from IPCC AR5 WGI TS). In addition you should be care of the historical period which last year should be 2005 (31 Dec 2005), because the projection after 1 Jan 2006 are wholly under climate change scenarios.

Dear MuhammadReza Tabatabaei ,

As per your answer kindly let me know when i want to perform bias correction fro my future data it would be based upon my observation as well as baseline data. The baseline data should also be used from 1986-2005 ? For one station is data is available since 1990 so is 15 year period suitable for the bias correction of future data for that particular station.

My first question is about using different observation periods for climatic stations in watershed in SWAT Model.

Dear Saima,

Regarding your first question you can use one common, overlapping period (1990-2015) for both stations or you can use the data of one station for the first 5 years (1985-1989) and the data of two stations for the remaining 25 years.

Regarding your second question, there is no standard period required for bias correction, although often climatic periods of 30 years are used. When applying bias correction (also when using a period of 15 years) it is important to split the data set in two parts: one for calibration and one for validation.

Dear Saima,

In addition to Martijn's answer, what I understand from SWAT model is that it will take closest meteorological stations from the center of each sub-catchment as the model forcing. Hence, your SWAT simulation can be mainly driven by one meteorological station depending on their proximity from sub-catchments. In addition to that, you should provide meteorological time series with an equal time length, otherwise the SWAT system will reject your simulation automatically (with a pop up warning on the screen).

For the second question, I agree with the previous answers. You may use your 20-years observed time series to perform bias correction. However, I will do what Martijn has said by means splitting the time series into two division for calibration and validation.

Good luck!