Has the way objects are connected changed in SWAT Editor 2.0.1?
Dennis Trolle
Hi all,
We have been experimenting for some time with SWAT+ 60.5 (with SWAT+ Editor 1.2.3 and QSWAT+ 1.4.0), and calibration using SWAT+ Toolbox – and we have been able to obtain quite good results in terms of hydrology.
We are now experimenting with the latest release of SWAT+ 60.5.2, including SWAT+ Editor 2.0.1 and QSWAT+ 2.0.5. Across the different model setups that we are running, we now seem to have the same difficulty in reproducing hydrographs. We all get relatively large shoulders on the recession curves, which do not match with observations. We have tried manipulating hard with groundwater (and many other) parameters without success. I have been reading through the commit messages between v60.5 and v60.5.2, as well as the fortran code itself, but I cannot find what could be the cause of this effect.
I then tried using the latest SWAT+ code (60.5.2) for a project set up using SWAT+ Editor 1.2.3 (and QSWAT+ 1.4.0), and then the results were good and close to those achieved when using v60.5.
If we try setting up a new project using SWAT+ Editor 2.0.1 and QSWAT+ 2.0.5, and calibrate through SWAT+ Toolbox, we get these large shoulders on the recession curves (and calibration does not help much). If I use the SWAT+ code version 60.5 for a project created using SWAT+ Editor 2.0.1 and QSWAT+ 2.0.5, the results are equally bad. I have attached a figure to illustrate this point.
This makes me wonder if the way the hydrological connections are made between spatial objects have changed between SWAT+ Editor v1.2.3 and v2.0.1 (if so, which change could be causing the effect we see)?
Has anyone experienced something similar (and does anyone know why the behavior of the model has changed)?
I tried looking for clues in the latest commit messages for SWAT+ Editor, but without success. Any hints will be greatly appreciated.
Thanks
Dennis
