Forest Disturbance Alters Water Yield but Not Catchment Drainage Behaviour

Abstract

Forested headwater watersheds that share climate and broadly similar lithology are often assumed to behave hydrologically alike, and forest-disturbance studies typically emphasize changes in annual water yield. To test this assumption, we analysed daily streamflow records spanning approximately 40–88 years from 14 independent headwater watersheds at the Coweeta Hydrologic Laboratory. Recession constant k ranged from 11.2 to 38.7 days, and baseflow index BFI ranged from 0.558 to 0.820. WS27, WS36 and WS37, whose divides exceed approximately 1450 m elevation, formed a distinct short-k/low-BFI cluster, whereas we detected no difference between the seven reference and seven treated watersheds representing disturbed forests (Mann–Whitney U = 20, exact two-sided p = 0.62 for both k and BFI). The 1965–2022 WS7–WS2 record spans 58 water years, of which 56 met the paired completeness criterion. The 1977 WS7 clear-cut increased annual yield relative to the pre-cut calibration by about 260 mm in the first post-cut year and by 68 mm year−1 on average over 1977–1989; after 1990 yield was no longer elevated (slightly below the WS2 calibration, −29 mm year−1). In contrast, under the primary strict-decline recession definition, k showed no detectable reference-adjusted step change relative to WS2. The contrasts were +0.7 days (95% interval−1.1 to 2.5) post-cut and −0.8 days (−1.9 to 0.3) in the recovered era. In a sensitivity analysis, recession runs bridged 1- or 2-day plateaus at the archive's resolution of 0.01 cubic feet per second per square mile (CSM). The post-cut contrast was +8% to +11% of the pre-cut value (intervals excluding zero). Thus, a modest transient lengthening of k is not excluded. BFI rose by 0.014 (0.003–0.028) relative to WS2 during the post-cut era (0.012, 0.001–0.024, relative to WS18) and was no longer distinguishable from its pre-cut relation in the recovered era (−0.006, −0.013 to 0.002, relative to WS2). These results identify k and BFI as persistent drainage signatures at Coweeta. Although they do not directly measure storage volume or critical-zone thickness, these metrics can screen watersheds for contrasting recession and low-flow behaviour that persists through forest disturbance.