Impact of antecedent conditions on simulations of a flood in a mountain headwater basin

Abstract

A devastating flood struck Southern Alberta in late June 2013, with much of its streamflow generation in the Front Ranges of the Rocky Mountains, west of Calgary. To better understand streamflow generation processes and their sensitivity to initial conditions, a physically based hydrological model was developed using the Cold Regions Hydrological Modelling platform (CRHM) to simulate the flood for the Marmot Creek Research Basin (~9.4 km2). The modular model includes major cold and warm season hydrological processes including snow redistribution, sublimation, melt, runoff over frozen and unfrozen soils, evapotranspiration, subsurface runoff on hillslopes, groundwater recharge and discharge and streamflow routing. Uncalibrated simulations were conducted for eight hydrological years and generally matched streamflow observations well, with a NRMSD of 52%, small model bias (−3%) and a Nash–Sutcliffe efficiency (NSE) of 0.71. The model was then used to diagnose the responses of hydrological processes in 2013 flood from different ecozones in Marmot Creek: alpine, treeline, montane forest and large and small forest clearings to better understand spatial variations in the flood runoff generation mechanisms. To examine the sensitivity to antecedent conditions, ‘virtual’ flood simulations were conducted using a week (17 to 24 June 2013) of flood meteorology imposed on the meteorology of the same period in other years (2005 to 2012), or switched with the meteorology of one week in different months (May to July) of 2013. Sensitivity to changing precipitation and land cover was assessed by varying the precipitation amount during the flood and forest cover and soil storage capacity in forest ecozone. The results show that runoff efficiency increases rapidly with antecedent snowpack and soil moisture storage with the highest runoff response to rainfall from locations in the basin where there are recently melted or actively melting snowpacks and resulting high soil moisture or frozen soils. The impact of forest canopy on flooding is negligible, but flood peak doubles if forest canopy removal is accompanied by 50% reduction in water storage capacity in the basin. Copyright © 2016 John Wiley & Sons, Ltd.

The cold rain‐on‐snow event of June 2013 in the Canadian Rockies — characteristics and diagnosis

“…The study was conducted in the upper elevations of Marmot Creek Research Basin (MCRB) (50°57′N, 115°09′W) in the Kananaskis Valley, Alberta, Canada, located in the Front Ranges of the Canadian Rocky Mountains (Figure ; Fang and Pomeroy, ). Elevation ranges from 1590 m.a.s.l at the Marmot Creek outlet to 2829 m.a.s.l at the summit of Mount Allan.…”

“…Over 3 days in late June 2013, 250 mm of precipitation fell on the partially snow‐covered and heavily instrumented Marmot Creek Research Basin, Canadian Rockies, contributing to the largest recorded flood in the region, the destruction of most gauging stations in the research basin and the most expensive natural disaster in Canadian history (Pomeroy et al , ). The event hydrometeorology has been described in detail by Liu et al (), the historical hydrological context of this flood described by Whitfield and Pomeroy () and the impact of antecedent conditions examined by Fang and Pomeroy (); these papers and the overview by Pomeroy et al () all suggest that snowmelt, including rain‐on‐snow (ROS), was an important contributor to the flood in addition to heavy rainfall. The remarkable celerity and synchrony of translation of precipitation into river discharge was noted by Pomeroy et al ().…”

Hydrological and temperature variations between 1900 and 2016 in the Catskill Mountains, New York, USA

“…Regardless of the uniqueness of the recent warm season precipitation record, these and previous results (Matonse and Frei, ; Frei and Kelly‐Voicu, ) suggest that seasonal mean conditions, for both precipitation and streamflow, have been unusually high since the late 1990s; and that the recent increase in seasonal extremes is more pronounced in the streamflow than in the precipitation record. This is consistent with the understanding that the magnitudes of extreme streamflow events are strongly correlated to antecedent conditions as well as to the magnitudes of precipitation events (Lumia et al ., ; Ivancic and Shaw, ; Fang and Pomeroy, ). In other words, during generally wetter conditions, when soils are on average wetter, streamflow will respond more dramatically to warm season precipitation events.…”

Spatio‐temporal aspects of snowpack runoff formation during rain on snow

“…Although on the point scale complex snow cover processes and uncertainties in meteorological forecast make it difficult to accurately predict snowpack runoff, assessing ROS on the catchment scale additionally comprises to deal with the spatial heterogeneity of snow cover properties and spatially variable meteorological inputs that influence both snowmelt and hydrological processes (Westrick & Mass, ). For example, high antecedent soil moisture is often observed during spring snowmelt conditions (Allan & Roulet, ; Fang & Pomeroy, ; Kampf, Markus, Heath, & Moore, ; Webb, Fassnacht, & Gooseff, ; Wever, Comola, Bavay, & Lehning, ) and can augment catchment runoff significantly. Preferential flow of liquid water through snow can have a distinct impact on timing and amount of snowpack runoff and has been examined using dye tracers (Schneebeli, ; Williams, Erickson, & Petrzelka, ; Würzer, Wever, Juras, Lehning, & Jonas, ), radar measurements (Albert, Koh, & Perron, ), temperature investigations (Conway & Benedict, ), and by measuring the spatial variability of snowpack runoff (Kattelmann, ; Marsh, ; Marsh & Pomeroy, ; Marsh & Woo, ).…”