gwflow baseflow discharge (gw_sw) not responding to recharge or aquifer parameters — SWAT+ rev. 62

[sharafat hussain]

Hi all,

I'm running a SWAT+ model with the gwflow module (rev. 62, SWAT+ Editor 4.0.1) for a ~297 km² watershed in South Korea, and comparing simulated streamflow against a real daily gauge record (2020–2024). Wet-season peaks match observed flow closely (PBIAS ≈ -3 to -8%), but dry-season flow is consistently and severely underestimated (PBIAS ≈ -76 to -81%) across every test below.

Looking at gwflow_basin_wb_mon.txt for 2023–2024, I see:

• recharge varies seasonally as expected (4.3–28.3 mm/month, tracking monsoon precipitation)
• gw_sw (groundwater discharge to streams) stays nearly flat regardless: -6.96 to -8.61 mm/month, showing essentially no response to the large swings in recharge
• dvol (storage change) is positive in nearly every month through 2023–2024, i.e. the aquifer keeps gaining storage without leveling into a seasonal equilibrium, even ~11 years into the simulation (4-year synthetic warm-up + observed record)

I've tested the following, each confirmed to have actually propagated into the run:

• ESCO 0.95 → 0.75, LAT_TTIME 0 → 60 (all HRUs)
• streambed_K 0.005 → 0.05 m/day (via Connections → Groundwater Flow → Zones)
• aquifer thickness 50m (a flat, apparently default value in cells.gw, inconsistent with my actual top_elev/bot_elev rasters which give a uniform 10m) → corrected to 10m
• forest CN2 lookup wood_f → wood_g for all forest land uses (85 of 101 HRUs)

None of these produced a meaningful change in the dry-season deficit or in the flatness of gw_sw. Aquifer K is otherwise a trusted value (1.0 m/day, uniform) carried over from an existing calibrated MODFLOW-NWT model of the same watershed.

Has anyone seen gw_sw behave this way — essentially decoupled from recharge/storage — in gwflow? Is there a setting I might be missing that would cause groundwater-to-stream discharge to be capped or otherwise insensitive to the aquifer's actual state? Happy to share full input files/output if useful.

Thanks in advance.

[Zayed Bin Saif]

Hi Sharafat,

Did you calibrate the model? Also, can you please share the other calibration matrices (such as NSE and KGE)?

I suspect that, during the dry season, flow from other sources might be added to the channel. Also, how many gauging stations have you selected for the comparison?

Thanks.

Best regards,

Zayed Bin Saif

MS Student 

Dept. of EES

University of Kentucky

--
You received this message because you are subscribed to the Google Groups "SWAT+ Editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swatplus-edit...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/swatplus-editor/88e09d19-546b-4b43-a12f-57319db048b5n%40googlegroups.com.

--

Zayed Bin Saif.
Sub-Divisional Engineer

Joint Rivers Commission, Bangladesh.