Hydrologic models often require correct estimates of surface macro-depressional storage to accurately simulate rainfall–runoff processes. Traditionally, depression storage is determined through model calibration or lumped with soil storage components or on an ad hoc basis. This paper investigates a holistic approach for estimating surface depressional storage capacity (DSC) in watersheds using digital elevation models (DEMs). The methodology includes implementing a lumped DSC model to extract geometric properties of storage elements from DEMs of varying grid resolutions and employing a consistency zone criterion to quantify the representative DSC of an isolated watershed. DSC obtained using the consistency zone approach is compared to DSC estimated by “brute force” (BF) optimization method. The BF procedure estimates optimal DSC by calibrating DRAINMOD, a quasi-process based hydrologic model, with observed streamflow under different climatic conditions. Both methods are applied to determine the DSC for relatively low-gradient coastal plain watersheds on forested landscape with slopes less than 3%. Results show robustness of the consistency zone approach for estimating depression storage. To test the adequacy of the calculated DSC values obtained, both methods are applied in DRAINMOD to predict the daily watershed flow rates. Comparison between observed and simulated streamflow reveals a marginal difference in performance between BF optimization and consistency zone estimated DSCs during wet periods, but the latter performed relatively better in dry periods. DSC is found to be dependent on seasonal antecedent moisture conditions on surface topography. The new methodology is beneficial in situations where data on depressional storage is unavailable for calibrating models requiring this input parameter. Copyright © 2012 John Wiley & Sons, Ltd.