Subject: Clibration result interpretation Posted: 9/11/2017 Viewed: 10111 times
Please help me. I am modelling 114 catchments using rainfall soil moisture method. I have put in the land area and also the climate data (prep, temp, relative humidity and left wind as the default) I have also input 13 rivers to which the catchments are linked. I went ahead to join all the rivers so that they flow into each other. Is that a mistake. I have also input naturalized stream flow data into all the 114 Quaternary catchments using the stream flow gauge object. Now I try calibrating a single branch. and the results I get is quite confusing. I don't know what I should be expecting. Please help me!!
I get 3 results as follows. I have just transposed them so you will see clearly
Streamflow (below node or reach listed) (Million Cubic Meter)
127 \
Oli River
Inflow 127 \ S11B (gauge)
January 432.8 1.8
February 667.4 1.7
March 446.1 1.8
April 286.6 1.8
May 120.7 1.8
June 71.5 1.8
July 70.1 1.8
August 59.7 1.8
September 95.2 1.8
October 116.6 1.8
November 216.6 1.8
December 360.4 1.8
Sum 2943.8 21.8
Streamflow Relative to Gauge (%) (Percent)
127 \ Oli River Inflow
January 6.28
February 11.54
March 6.25
April 4.34
May 1.86
June 1.22
July 1.12
August 0.95
September 1.59
October 1.61
November 3.66
December 5.20
Ms. Stephanie Galaitsi
Subject: Re: Clibration result interpretation Posted: 9/13/2017 Viewed: 10088 times
Hi Darlington,
Without knowing your system, I can't answer whether you should join your rivers or not. The goal is to make your model as close to the actual basin as possible, in terms of river confluences and streamflow gauge locations. Did you use the gauges as "pour points" to determine the size of each basin feeding into them?
On to your question:
Result #1
127 \
Oli River
Inflow 127 \ S11B (gauge)
[These values should show the historic gauge data that you have plugged into your model, using available data from streamflow monitoring stations. I'm hoping you have values for multiple years?]
Result #2
Streamflow Relative to Gauge (%) (Percent)
127 \ Oli River Inflow
[This result compares simulated and observed flows. For values below zero, the gauge value, the observed volume recorded at the monitoring station, is higher than the value that your WEAP model is producing for that time step. For values below zero, your WEAP model is producing more water than what was historically observed. The better calibrated your model is, the closer these percentage values will be to zero].
For calibration purposes, we also recommend calculating and reporting your model's Nash Sutcliffe value. You can read about it here:
Subject: Re: Clibration result interpretation Posted: 9/14/2017 Viewed: 10052 times
Thanks so much for the clarifications. Now I know what to look for. I am not clear with the term 'pour point' can you please explain. What I did was to put the gauges just down stream of the Quaternary catchments. I'm not sure if that is what you mean?
Ms. Stephanie Galaitsi
Subject: Re: Clibration result interpretation Posted: 9/14/2017 Viewed: 10044 times
Hi Darlington,
The idea of a pour point is an important part of the catchment modeling in WEAP. A pour point can be any point of interest in your model where you want the flow in the river to be accurate. It might be right below a reservoir, it might be at the intake for an urban demand center, and it might be a streamflow gauge that you will use to calibrate your historic model. In that sense, pour points are determined by either the specific questions that your model is trying to answer, or the gauge data available.
Once you have the pour point - that place where the flow should be accurate - you design the catchments upstream to encompass the entire area that will contribute flow to that point. For example, if I have a pour point on the river, and in my natural system, 350 km2 contribute flow to that point, my modeled catchment area should be 350km2 (and if you're modeling different land use types, they should be the land use types present in that area). If there is a mismatch, that your model will generate either too much flow or too little flow. You can also have subcatchments within catchments, as long as they all converge to the pour point under examination.
The next version of WEAP will have automatic catchment delineation based on the pourpoints that the users choose. Then you can bypass the GIS work altogether.
