Sediment calibration in SWAT-CUP

[Abhinav Gupta]

Hello Users

I have to calibrate for sediment in my watershed. I followed the general advice of calibrating flow first. But after that when I calibrate sediment I am not able to get a good model (NS = 1.49). I followed the  suggestions in SWAT-CUP manual for sediment calibration, but I am not sure if I have implemented everything correctly. So, can someone help me to understand what is it that I am doing wrong. Here is what I have done:

(1) I got a good calibration for flow (NS = 0.71).
(2) Now I removed FLOW_OUT data from observed_rch.txt and observed.txt and added sediment load to these files.
(3) Then, I updated the range of the parameters for flow suggested by SWAT-CUP in my last iteration and added some new parameters in PAR_INF.txt, which affect only sediment.
(4) Then, I ran 500 simulation.

Also, some minor changes in SWAT-CUP input files corresponding to SED_OUT were made. These are the only steps that I have taken. I don't know what am I missing?

Best
Abhinav Gupta.

[Karim]

Basically correct steps. No need to remove the flow from observation files. You can still slightly modify flow parameters to get a better sediment results. But Sediment is in general hard to calibrate, because of large measurement errors and the fact that usually grab samples are taken. Also, what swat gives and what is measured may be different things.

Karim

--
You received this message because you are subscribed to the Google Groups "SWAT-CUP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swat-cup+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Abhinav Gupta]

Thank you, Dr. Karim! I don't understand the last sentence. I had sediment concentration values which I converted to load by multiplying concentration with flow and time. does not it what SWAT gives as SED_OUT?
Also, does removing flow would affect my calibration anyway because I am not calibrating flow?

[Abbaspour, Karim]

How are the sediments measurements taken and how often?

Flow parameters also influence sediment. You should allow slight recalibration of flow parameter ranges in order to get better sediment simulation.

Karim

[Abhinav Gupta]

Sediment measurements are taken as grab samples and there are 81 of sediment concentration in 10 years (1996-2004).

[Abbaspour, Karim]

And is this what swat simulates?

Karim

 

 

From: swat...@googlegroups.com [mailto:swat...@googlegroups.com] On Behalf Of Abhinav Gupta
Sent: Friday, March 24, 2017 10:22 PM
To: SWAT-CUP
Subject: Re: Sediment calibration in SWAT-CUP

 

Sediment measurements are taken as grab samples and there are 81 of sediment concentration in 10 years (1996-2004).

--

[Abhinav Gupta]

Yes, according to SWAT output file manual it generates daily sediment load in tons.

On Thursday, March 23, 2017 at 9:17:53 PM UTC-4, Abhinav Gupta wrote:

[Karim]

Well swat gives you a daily average. Grab samples are not daily averages, they are usually taken at or after a rainfall event, usually when concentrations are high. Also, it depends at what part of the river they have been taken from. Perhaps at a hight turbulent place in the river?  

Karim

[Abhinav Gupta]

So, I could figure out that my measurements are only suspended sediment; however, SWAT output.rch gives total sediment loads.
However, total suspended sediment (TSS) is contained in output.sed, not in output.rch file; but unfortunately SWAT-CUP can not extract data from output.sed.
So, I was wondering if I can manually add the TSS column from output.sed into output.rch file contained in main SWAT-CUP folder and calibrate  using output_rch.txt.

Best
Abhinav Gupta.

On Thursday, March 23, 2017 at 9:17:53 PM UTC-4, Abhinav Gupta wrote:

[Aslam Hanief]

TSS is column 12 in the output.rch file. Look at the values, they are the same. 

--
You received this message because you are subscribed to the Google Groups "SWAT-CUP" group.

To unsubscribe from this group and stop receiving emails from it, send an email to swat-cup+unsubscribe@googlegroups.com.

[Abhinav Gupta]

Column 12 is sediment concentration. It seems like total sediment concentration. SWAT output file manual says that SEDCONC is 'Concentration of sediment in reach during time step (mg/L)'. But it is not mentioned that it is suspended sediment. Is there is any logic to why it is suspended sediment only.

On Thursday, March 23, 2017 at 9:17:53 PM UTC-4, Abhinav Gupta wrote:

[Aditya Nilwar]

Hi Abhinav 

Can you plz share how you calibrate TSS in SWAT CUP?

[LEGEND LEGEND]

Abhinav

 I have sediment data daily basis (Grab samples) Total concentration in PPM and Concentration >75 μm in PPM . I don't know the meaning of  >75 μm sediment concentration please tell me about that if u have idea. Which data should i use for calibration? Also tell me how to convert Total concentration in PPM to ANOTHER UNIT THAT IS USED for calibration?/

On Sunday, March 26, 2017 at 8:58:47 PM UTC+5:30, Abhinav Gupta wrote:

[Karim Abbaspour]

>75 is the size of the grains. Means it does not pass through a 75 micro meter sieve. To go from concentration to load, you have to multiply by the flow. Measurement error for sediment is often quite high, so is the simulation error. Therefore, I would not take sediment calibration too seriously! I would just look at the annual load and make sure it is within the ballpark.

Karim

--
You received this message because you are subscribed to the Google Groups "SWAT-CUP" group.

To unsubscribe from this group and stop receiving emails from it, send an email to swat-cup+u...@googlegroups.com.

[LEGEND LEGEND]

Hello all

Warm greetings.

After writing back the calibrated parameters in Arc SWAT, i got same results for flow ie Arc SWAT simulated flow matches with SWAT CUP simulated Flow but, it doesn't match with Sediment yield. What may be the reason? I tried a lot to find it but couldn't get the solution. It took my whole 1 month. (I Want to use my model further for Best Management Practices). Monthly Sediment yield From Arc SWAT is much higher than that in SWAT CUP with same parameter values assigned.

On Sunday, March 26, 2017 at 8:58:47 PM UTC+5:30, Abhinav Gupta wrote:

[LEGEND LEGEND]

Hello all

Warm greetings.

After writing back the calibrated parameters in Arc SWAT, i got same results for flow ie Arc SWAT simulated flow matches with SWAT CUP simulated Flow but, it doesn't match with Sediment yield. What may be the reason? I tried a lot to find it but couldn't get the solution. It took my whole 1 month. (I Want to use my model further for Best Management Practices). Monthly Sediment yield From Arc SWAT is much higher than that in SWAT CUP with same parameter values assigned.

[Karim Abbaspour]

Probably different swat versions.

There is no reason to put parameters back into ArcSwat. Once you calibrated with swatcup, you can do everything in that folder. Your solution is the range of parameters and you cannot put this into ArcSwat. After calibration, you always have to propagate the uncertainty (parameter range) and you can do this easily in the swatcup folder. A BMP scenario changes a couple of parameters in some swat files. You can copy those files only into swatcup folder and perform an iteration to see the results in swatcup.

Hope this is clear.

Karim

On Jan 20, 2018, at 13:38, LEGEND LEGEND <neupane...@gmail.com> wrote:

Hello all

Warm greetings.

After writing back the calibrated parameters in Arc SWAT, i got same results for flow ie Arc SWAT simulated flow matches with SWAT CUP simulated Flow but, it doesn't match with Sediment yield. What may be the reason? I tried a lot to find it but couldn't get the solution. It took my whole 1 month. (I Want to use my model further for Best Management Practices). Monthly Sediment yield From Arc SWAT is much higher than that in SWAT CUP with same parameter values assigned.

On Friday, January 19, 2018 at 5:30:11 PM UTC+5:30, LEGEND LEGEND wrote:

Hello all

Warm greeting

After writing back the calibrated parameters in Arc SWAT, i got same results for flow ie Arc SWAT simulated flow matches with SWAT CUP simulated Flow but, it doesn't match with Sediment yield. What may be the reason? I tried a lot to find it but couldn't get the solution. It took my whole 1 month. (I Want to use my model further for Best Management Practices). Monthly Sediment yield From Arc SWAT is much higher than that in SWAT CUP with same parameter values assigned.

On Sunday, March 26, 2017 at 8:58:47 PM UTC+5:30, Abhinav Gupta wrote:

Column 12 is sediment concentration. It seems like total sediment concentration. SWAT output file manual says that SEDCONC is 'Concentration of sediment in reach during time step (mg/L)'. But it is not mentioned that it is suspended sediment. Is there is any logic to why it is suspended sediment only.

On Thursday, March 23, 2017 at 9:17:53 PM UTC-4, Abhinav Gupta wrote:

Hello Users

I have to calibrate for sediment in my watershed. I followed the general advice of calibrating flow first. But after that when I calibrate sediment I am not able to get a good model (NS = 1.49). I followed the  suggestions in SWAT-CUP manual for sediment calibration, but I am not sure if I have implemented everything correctly. So, can someone help me to understand what is it that I am doing wrong. Here is what I have done:

(1) I got a good calibration for flow (NS = 0.71).
(2) Now I removed FLOW_OUT data from observed_rch.txt and observed.txt and added sediment load to these files.
(3) Then, I updated the range of the parameters for flow suggested by SWAT-CUP in my last iteration and added some new parameters in PAR_INF.txt, which affect only sediment.
(4) Then, I ran 500 simulation.

Also, some minor changes in SWAT-CUP input files corresponding to SED_OUT were made. These are the only steps that I have taken. I don't know what am I missing?

Best
Abhinav Gupta.