Research papersPaired air-water annual temperature patterns reveal hydrogeological controls on stream thermal regimes at watershed to continental scales
Introduction
As climate change and other anthropogenic alterations to watersheds transform the natural thermal regime of streams and rivers (Bassar et al., 2016, Isaak et al., 2012, Kaushal et al., 2010, Kędra and Wiejaczka, 2018), understanding the sensitivity and vulnerability of stream segments to these changes is increasingly imperative for long-term ecological management. Temperature is one of the most important properties controlling the water quality of streams (Caissie, 2006). Almost all physical, chemical, and biological processes of the stream corridor are influenced by temperature, including fish development and metabolism, dissolved oxygen concentration, biogeochemical cycling, and organic matter decomposition. Absent hydrologic alteration, channel water temperature is driven by meteorological and hydrogeological factors, such as incoming solar radiation, outgoing longwave radiation, air temperature, wind speed, humidity, stream channel dimensions (depth, width, and flow volume), and groundwater (GW) inputs (Caissie, 2006, Cluis, 1972, Westhoff et al., 2007). Because of its effect on ecosystem processes, stream temperature has received abundant attention in the scientific literature for more than a half-century and numerous modeling approaches have been used to predict different facets of its regime (key reviews include Anderson, 2005, Benyahya et al., 2007a, Caissie, 2006, Gallice et al., 2015, Kurylyk et al., 2019, Webb et al., 2008). Short-term air-water temperature relations have been used to map spatially variable stream water thermal sensitivity in summer (Kelleher et al., 2012). However, longer-term annual water temperature patterns, and how these relate to local air temperature patterns and watershed processes, have been underexplored.
Stream temperature models fall into deterministic or statistical groups (Benyahya et al., 2007a, Caissie, 2006). Deterministic models are based on the balance of energy (heat) and mass (flow) fluxes in a water body (Boyd and Kasper, 2003, Glose et al., 2017). These models are best for conducting impact studies that assess changes to one or more components of the stream heat budget or when exploring changes in temperature at multiple spatial scales (Benyahya et al., 2007a, Caissie, 2006, Westhoff et al., 2007). However, total heat budget models are complex, computationally expensive, and require numerous hydrological, physiographic, and meteorological inputs that may be excluded from typical measurement protocols, poorly defined at the spatial scale needed for ecological applications, or difficult to quantify. In contrast, statistical models are computationally simpler with minimal data requirements (Benyahya et al., 2007a) facilitating prediction at ecologically relevant spatial grains and extents. However, statistical models currently lack a clear understanding of the relationships between derived model coefficients and important watershed processes potentially limiting their utility. Further, unsampled spatial heterogeneity can lead to overly simplistic predictions at multiple scales. For instance, the development of national-scale models, such as The National Water Model (NOAA, 2019), requires understanding of intra- and inter-regional spatial variation for meaningful downscaled predictions.
Statistical models typically use stochastic methods to assess relationships between water and air temperatures for small time-steps (e.g., daily). These models result in the derivation of a long-term periodic component of water temperature data, which is commonly assumed to be temporally stable (Caissie et al., 1998). However, interannual and long-term trends in water temperature may be driven by temporally variable watershed processes, including solar radiation, GW discharge dynamics from adjacent aquifers and temporal runoff patterns. While tracing heat signatures has long been used to monitor the activity of these and other contributing factors (Anderson, 2005, Constantz, 2008, Halloran et al., 2016), few studies have explored their influence on patterns and properties of the long-term water temperature signal (i.e., mean, phase, and amplitude). Together, these properties can indicate the presence of stream thermal inertia (Letcher et al., 2016), influence of shallow (~upper 6 m) versus deeper GW discharge (Briggs et al., 2018b), riparian shading (Fabris et al., 2018, Johnson and Jones, 2000, Wawrzyniak et al., 2017), or dam operations (Rounds, 2007). However, it is unclear whether these observations are applicable across locations and climates and if annual stream temperature signals are relatively stable over time. Further, air temperature is often assumed to be a dominant control on stream temperature at seasonal timescales, but recent research indicates uncoupling of the annual signal amplitude and phase relations between air and water temperature may be used to infer other watershed controls on the stream heat budget (Briggs et al., 2018a).
Here, we assess the utility of simple statistical modeling approaches using multi-year paired air-water temperature data for developing ecologically relevant thermal metrics, and to evaluate how these diagnostic metrics vary across multiple spatial scales. The specific objectives are to: (1) compare annual air and water temperature signal parameters and combined air-water metrics across watershed, regional, and national spatial scales; and (2) provide guidance for the interpretation of three paired air-water annual thermal metrics (mean ratio, phase lag, and amplitude ratio) derived from sine-wave linear regressions with respect to physical watershed processes. These objectives are addressed using paired air and water temperature data collected at relatively high spatial resolution in mountain headwater streams from two different climatic regions of the United States (U.S.) (Pacific Northwest and Mid-Atlantic), along with data from streams of generally larger size distributed across the contiguous U.S. The use of the latter, national dataset, allows for a comparison between the paired air-water annual temperature signals method presented here and another multi-year thermal classification method that used the same sites (Maheu et al., 2016). This previously developed stream thermal regime classification system, using identified key environmental drivers, was based exclusively on water temperature data. Their results showed that sites could be clustered based on differences in the annual water temperature mean, amplitude, and phase parameters, and that these clusters could be predicted by air temperature and flow characteristics. We hypothesize that the inclusion of the comparison of annual water temperature patterns to local annual air temperature patterns in the present analysis will improve the ability to predict processes such as GW exchange and riparian vegetation characteristics that may vary across steep gradients or spatial discontinuities.
Section snippets
Data and methods
In this section, we describe the site characteristics of the regional and national datasets used in this study and methods for analyzing annual air and water temperature sine-wave signals with linear regression.
Results
In summary, OESF and SHEN thermal regimes were dramatically different when annual air and water temperature signal parameters were compared separately. However, when air and water temperature signals were combined (i.e., MR, Δɸ, and AR), OESF and SHEN sites generally exhibited similar patterns of shallow GW discharge dominance, as indicated by a negative Δɸ-AR relation. The generally larger USGS sites exhibited greater spatial variation in the annual air and water temperature parameters among
Discussion
In this study, paired air and water annual temperature signals were analyzed and summarized at watershed, regional, and national spatial scales. These annual patterns are also discussed in the context of upstream physical watershed characteristics. Our results demonstrate the utility of paired air-water annual temperature data for inferring hydrogeological processes in streams. Specifically, we show that (1) sine-wave regression yields useful statistical parameters for interpreting GW influence
Conclusions
Paired air-water annual temperature signals are a promising tool for efficiently diagnosing GW influence and other major controls on stream water thermal regimes (e.g., dam operation, riparian shade, etc.) from readily available data sources. Previous thermal regime classifications based on seasonal water temperatures alone tended to cluster when plotted in air-water annual temperature metric space in expected ways, however, including the paired air-water annual signal analysis adds critical
CRediT authorship contribution statement
Zachary C. Johnson: Conceptualization, Methodology, Writing - review & editing, Data curation. Brittany G. Johnson: Conceptualization, Methodology, Writing - review & editing. Martin A. Briggs: Conceptualization, Writing - review & editing. Warren D. Devine: Data curation, Writing - review & editing. Craig D. Snyder: Conceptualization, Data curation, Writing - review & editing. Nathaniel P. Hitt: Conceptualization, Data curation, Writing - review & editing. Danielle K. Hare: Data curation,
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
Funding for this work was provided by the University of Washington School of Environmental and Forest Sciences and the U.S. Geological Survey Toxic Substances Hydrology Program. Additional information and data are available as Appendices to this manuscript, which can be found in the Supplementary information. We thank Andy Gendaszek, Sandra Cooper, and two anonymous reviewers for helpful comments and suggestions. Any use of trade, firm, or product names is for descriptive purposes only and does
References (104)
- et al.
Inferring watershed hydraulics and cold-water habitat persistence using multi-year air and stream temperature signals
Sci. Total Environ.
(2018) - et al.
Modelling of maximum daily water temperatures in a small stream using air temperatures
J. Hydrol.
(2001) - et al.
Spatial distribution of thermal refuges analysed in relation to riverscape hydromorphology using airborne thermal infrared imagery
Remote Sens. Environ.
(2015) - et al.
Stream temperature under contrasting riparian forest cover: understanding thermal dynamics and heat exchange processes
Sci. Total Environ.
(2018) Relative importance of spatial and temporal scales in a patchy environment
Theor. Popul. Biol.
(1992)- et al.
Stream heat budget modeling with HFLUX: Model development, evaluation, and applications across contrasting sites and seasons
Environ. Model. Softw.
(2017) - et al.
Heat as a tracer to quantify processes and properties in the vadose zone: a review
Earth-Sci. Rev.
(2016) Stream temperatures in an Alpine area
J. Hydrol.
(1971)- et al.
Climatic and dam-induced impacts on river water temperature: assessment and management implications
Sci. Total Environ.
(2018) - et al.
Monitoring riverine thermal regimes on stream networks: insights into spatial sampling designs from the Snoqualmie River
WA. Ecol. Indic.
(2018)