INTRODUCTION

Uncertainty in the representation of hydrological processes at local scales in Earth system models can affect the accuracy of the weather forecasts and climate projections. Water budget and surface energy in large basins, which are formed by small homogenous hydrological response units, are primarily affected by regional climatic teleconnection patterns. Small variations in climatic patterns can lead to large hydrological responses (Whitfield, 2001), and can affect hydrological predictability (Rasouli, Hsieh, & Cannon, 2012). Interactions between local and regional scale hydroclimatic fluxes, however, are not sufficiently resolved in the Earth system and climate models used for environmental change studies (Fan et al., 2019). Whitfield, Moore, Fleming, and Zawadzki (2010) studied the low frequency (e.g., multi-decadal) and high frequency (e.g., monthly) variations of climate and their associations with hydrological changes. A multi-decadal component of spring streamflow, for instance, is associated with variations in spring precipitation and air temperature driven by low frequency atmospheric circulations and sea level pressure (Boé & Habet, 2014). In snow-dominated mountains with shallow soils and high spatial heterogeneity, streamflow variations also depend on low frequency variations of groundwater. Because of the lower velocity of groundwater relative to surface runoff, it can take three to eight years to recycle and contribute to surface flows in mountainous areas, depending on the variability of meltwater from fresh snowpack (e.g., Plummer et al., 2001; Manning et al., 2012). The decomposition of local hydrological time series into its components can yield comparable scales that are needed to relate climate circulations with a low frequency and long cycles to hydrological fluxes with a high frequency and short durations. Understanding hydrological variations with an intermediate frequency (e.g., occurring every 3 – 8 years) remains challenging in mountainous regions with large geological heterogeneity as attribution of antecedent energy and moisture conditions in the previous years to seasonal snow and flow regimes in the following year cannot be easily monitored or estimated. Spatial and temporal variations of snowmelt, runoff, groundwater storage and flow, antecedent soil moisture, snow redistribution by blowing wind, and snow sublimation, are high, which makes modeling and predicting mountain hydrology uncertain (Lehning, Grünewald, & Schirmer, 2011). This becomes even more challenging when hydrological models are forced with atmospheric fluxes with uncertain measurements and large natural variations.
Reynolds Creek Experimental Watershed (RCEW), with a semiarid cool montane climate in Idaho, USA, is of specific scientific interest. For example, its critical zone responses to climate change have been monitored over three decades and subsequently modeled by many studies (e.g. Seyfried, Grant, Marks, Winstral, & McNamara, 2009; Reba et al., 2011a; Kumar, Wang, & Link, 2012; Marks, Winstral, Reba, Pomeroy, & Kumar, 2013; Rasouli, Pomeroy, & Marks, 2015). Reynolds Mountain East (hereafter, Reynolds Mountain), as a headwater basin within RCEW, has been widely investigated to understand hydrological changes in the mountains and often cited in the literature as a representative basin for semiarid snow dominated regions (e.g. Marks & Winstral, 2001; Reba et al., 2011a; Kumar, Marks, Dozier, Reba, & Winstral, 2013; Wang, Kumar, & Marks, 2013; Chen, Kumar, Wang, Winstral, & Marks, 2016). Reba, Marks, Winstral, Link, and Kumar (2011b) conducted a detailed study of the sensitivity of the snow cover energetics and classified hydrological simulations into eight categories based on annual and winter precipitation and snowpack. All eight categories in Reynolds Mountain, however, were temporally discontinuous. Dry and cool conditions, for instance, in 1985 and 2000, may not belong to the same teleconnection phase and may have different atmospheric driving mechanisms. Interactions between local hydrological dynamics and regional climatic teleconnection patterns, along with a physically based representation of these interactions in climate models, can improve the understanding of climate and hydrological processes at local scales (Prein et al., 2015) and the understanding of weather and climate extremes at regional scales (Langendijk et al., 2019). The large biases that weather and climate model outputs show against observations (Fowler, Blenkinsop, & Tebaldi, 2007) can be reduced by linking land surface processes and local atmospheric convection to climatic teleconnection patterns across a range of temporal and spatial scales.
The attribution of regional climate patterns over the preceding years to seasonal hydrological fluxes at local scales in the following year has not been sufficiently understood. Therefore, the linkage between local hydrological processes and regional climatic teleconnections patterns was explored in this study. The linkage between local hydrological variations and climatic teleconnection patterns, assessed over multiple hydroclimatic phases in this paper, can be missed when only seasonal precipitation, snowmelt runoff, and evapotranspiration are investigated. Misrepresenting low (e.g., multiple decades), intermediate (e.g., 3 – 8 years), and high (e.g., multiple months) frequency variations can result in uncertainties in climate projections and failure in both short and long-term hydrological predictions (Dutta & Maity, 2018). This study addresses the following research questions: how do local hydrological processes in a small basin relate to regional climatic patterns, and how does the spatial variability of hydrological fluxes in snow dominated mountain basins differ under different phases of atmospheric circulations?