Anyone else seeing a persistent gap when porting GOHFER-derived geometry into IMEX for production forecasting? On a 9,500‑ft Wolfcamp A lateral with 42 stages (2023), I can match the first 180 days only by inflating SRV perm to about 0.05 mD and enabling pressure‑dependent permeability, but the model then overpredicts months 6–12 even with time‑dependent conductivity and non‑Darcy; what transport/leakoff assumptions have given you stable forecasts post‑cleanup?
If you’re matching the first 180 d by cranking SRV k, try the opposite: keep SRV k/phi near core-scale and use a ‘sigma–FCD’ table in IMEX from GOHFER’s closure/width, then apply a time-dependent fracture-surface multiplier after about 180 d so late flow is area‑limited instead of k‑limited. Did you import the closure/width curve or just geometry?
Late overshoot screams an over‑connected SRV; keep SRV k near core‑scale and in IMEX damp fracture–matrix transfer beyond about 250–300 ft by cutting the dual‑porosity shape factor (or X/Y MULT about 0.3) in the outer SRV, and restrict PPD‑k to fracture cells only. On that 9,500‑ft Wolfcamp A with 42 stages, this preserves the 0–180 d match but steepens months 6–12 to reality; piggybacking on @ian_swift99, it’s the connectivity, not bulk SRV k. Can you try a distance‑weighted SIGMA and see if the month‑8–10 gap closes?