Infiltration into frozen soil: From core‐scale dynamics to hillslope‐scale connectivity

Abstract

Infiltration into frozen soil is a key hydrological process in cold regions. Although the mechanisms behind point‐scale infiltration into frozen soil are relatively well understood, questions remain about upscaling point‐scale results to estimate hillslope‐scale run‐off generation. Here, we tackle this question by combining laboratory, field, and modelling experiments. Six large (0.30‐m diameter by 0.35‐m deep) soil cores were extracted from an experimental hillslope on the Canadian Prairies. In the laboratory, we measured run‐off and infiltration rates of the cores for two antecedent moisture conditions under snowmelt rates and diurnal freeze–thaw conditions observed on the same hillslope. We combined the infiltration data with spatially variable data from the hillslope, to parameterise a surface run‐off redistribution model. We used the model to determine how spatial patterns of soil water content, snowpack water equivalent (SWE), and snowmelt rates affect the spatial variability of infiltration and hydrological connectivity over frozen soil. Our experiments showed that antecedent moisture conditions of the frozen soil affected infiltration rates by limiting the initial soil storage capacity and infiltration front penetration depth. However, shallow depths of infiltration and refreezing created saturated conditions at the surface for dry and wet antecedent conditions, resulting in similar final infiltration rates (0.3 mm hr−1). On the hillslope‐scale, the spatial variability of snowmelt rates controlled the development of hydrological connectivity during the 2014 spring melt, whereas SWE and antecedent soil moisture were unimportant. Geostatistical analysis showed that this was because SWE variability and antecedent moisture variability occurred at distances shorter than that of topographic variability, whereas melt variability occurred at distances longer than that of topographic variability. The importance of spatial controls will shift for differing locations and winter conditions. Overall, our results suggest that run‐off connectivity is determined by (a) a pre‐fill phase, during which a thin surface soil layer wets up, refreezes, and saturates, before infiltration excess run‐off is generated and (b) a subsequent fill‐and‐spill phase on the surface that drives hillslope‐scale run‐off.

Effects of preferential flow on snowmelt partitioning and groundwater recharge in frozen soils

“…Such an effect likely contributed to the sequential increase in runoff ratios during melt events at the Stauffer and Triple G sites (Table 3). These observations of the refreezing of infiltrated meltwater are generally consistent with the "pre-fill" phase of the conceptual model of runoff generation in frozen soil proposed by Appels et al (2018), in which the refreezing of initially infiltrated water is necessary to reduce infiltrability before runoff can occur. To the best of our knowledge, no other study has demonstrated the time sequencing of this infiltration-refreezing effect on enhancing subsequent runoff partitioning over multiple snowmelt events.…”

Evaluation of the Water Conservation Function of Different Forest Types in Northeastern China

“…Dunkerley et al have demonstrated that the more compact the soil, the smaller its porosity and the lower the steady infiltration rate [40]. With a higher aggregate content, the number of soil pores increases, resulting in improved soil ventilation and permeability and in enhanced soil stability, preventing the formation of soil surface crusts [41]. The aggregate content was highest in the soil from the L. gmelinii site, which therefore had the highest seepage stabilization rate.…”

“…The aggregate content was highest in the soil from the L. gmelinii site, which therefore had the highest seepage stabilization rate. demonstrated that the more compact the soil, the smaller its porosity and the lower the steady infiltration rate [41]. With a higher aggregate content, the number of soil pores increases, resulting in improved soil ventilation and permeability and in enhanced soil stability, preventing the formation of soil surface crusts [42].…”

Fifty years of recorded hillslope runoff on seasonally frozen ground: the Swift Current, Saskatchewan, Canada, dataset

“…Additional information on the generation of the 2 m and 0.25 m DEMs can be found in . The soils are Orthic Brown Chernozems (Ayres et al, 1985), specifically Swinton silt loam in the upper four-fifths of the hillslopes, where the solum is developed within loess, and Haverhill loam in the lower fifth of the hillslope, where past erosion has resulted in the solum extending into the underlying loamy till (McConkey et al, 1997). The surficial 0-15 cm soil layer consists of 50.4 % silt, 31.4 % sand, and 18.2 % clay, has a bulk density of 1.22 g cm −3 , and has a saturated hydraulic conductivity of 14.2 mm h −1 .…”

“…The year 2011 marked the final year of regular monitoring of the hillslopes by the Swift Current Research and Development Centre of Agriculture and Agri-food Canada. In 2013 and 2014, monitoring of runoff, snowpack characteristics, and soil moisture was undertaken in collaboration with the University of Saskatchewan; these data are presented separately (Appels et al, 2017;. If regular monitoring of the hillslopes resumes, the data repository will be updated accordingly.…”