Extended growing seasons and decreases in hydrologic connectivity indicate increasing water stress in humid, temperate forests
Introduction
Streamflow originating in forested regions is essential to support abundant and clean freshwater resources (Duan et al., 2018). Under ongoing global change, the forest hydrological cycle is intensifying, and in some cases the portion of precipitated water returning rapidly to the atmosphere is increasing, reducing downstream water availability. Increases in forest water use, or evapotranspiration (ET), have been linked to rising temperatures and subsequent increases in evaporative demand, largely driven by increasing vapor pressure deficit (VPD) (Frank et al., 2015; Niu et al., 2019; Novick et al., 2016; Wang et al., 2022). Additionally, increases in ET have been attributed to ecophysiological responses to changing climate, such as expanded growing seasons and subsequent vegetation growth, as well as long term changes in forest structures and functions (Frank et al., 2015; Gaertner et al., 2019; Hwang et al., 2018). In the humid, temperate forests of the eastern United States (US), species composition has shifted towards species with higher water use rates under a history of fire suppression, often called ‘mesophication’ (Caldwell et al., 2016). Despite identification of the drivers of forest water use, the extent that forest water use is increasing and how that translates to shifts in hydrologic partitioning within and across watersheds remains unclear.
Lateral hydrologic connectivity is an essential concept of hillslope hydrology describing the continuity of hydrologic flow paths along the hillslope-riparian-stream continuum (Detty and McGuire, 2010; Jencso et al., 2009; Jencso and McGlynn, 2011; McGuire and McDonnell, 2010). In this study, we specifically examine the connectivity of shallow subsurface flow accessible to vegetation along the hydrologic flow paths linking hillslope and riparian vegetation. Catchment-scale hydrologic connectivity is controlled by climate drivers, including precipitation patterns and water storage, and physical characteristics of watersheds, including topography, geology, and soil hydraulic properties (Ali and Roy, 2010; Hopp and McDonnell, 2009; Hwang et al., 2012; Shen et al., 2013). Soil moisture often follows the pattern of lateral flow along the hillslope gradient, with lower soil moisture found upslope and higher soil moisture in convergent and riparian areas (Ali and Roy, 2010; Fan, 2015; Tromp-van Meerveld and McDonnell, 2006). Vegetation patterns also organize along hydrologic flow paths according to their soil moisture requirements and dependence on the water and nutrients provided by lateral flows (Day et al., 1988; Hwang et al., 2009). Moist soils in convergent and riparian areas often support greater vegetation density composed of tree species with higher productivity and water use rates compared to upslope positions (Hwang et al., 2012, 2020; Mackay and Band, 1997; Tai et al., 2020). Therefore, lateral flow mediated by topography results in emergent patterns in vegetation structure and function at the watershed scale.
Both water and energy availability are important controls on vegetation function. Upslope forest water use and productivity are often limited by soil moisture due to topographic and edaphic factors that promote rapid drainage (Mackay and Band, 1997; Tai et al., 2020; Thompson et al., 2011). Energy is an important limiting factor at downslope positions with less exposed topography due to receiving less radiation (Tovar-Pescador et al., 2006). While eastern humid forests are typically considered energy-limited, vegetation organization along hydrologic flow paths in steep catchments indicates they are also shaped by water availability (Hwang et al., 2012). Consideration of the effects of topography mediating vegetation responses to changing climate is important to understand potential changes in lateral hydrologic connectivity and subsequent streamflow generation.
Vegetation water use is tightly coupled with photosynthesis, so vegetation patterns along hydrologic flow paths effectively represent long-term hydrologic partitioning between localized vegetation water use and shallow subsurface flow given topoclimate settings. In the southern Appalachians, shallow subsurface flow is a main source of sustained baseflow, even during dry periods (Hewlett and Hibbert, 1967; Hwang et al., 2009, 2012). Therefore, long-term trends in vegetation patterns at the watershed scale, often derived from remote sensing indices, can indicate changes in hydrologic connectivity, which are critical to predicting shifts in runoff generation dynamics (Ali and Roy, 2010; Hwang et al., 2012, 2020). For example, when ET increases upslope due to changes in vegetation dynamics (i.e., lengthened growing season, increasing biomass, expansion of species with high water use), less upslope subsidy is available to vegetation downslope and to supply runoff (Caldwell et al., 2016; Orth and Destouni, 2018). Hwang et al. (2020) used the remotely sensed normalized difference vegetation index (NDVI) to quantify relative change in vegetation density between upslope and downslope portions of headwater catchments, classified according to upslope contributing area, to infer long-term changes in connectivity of shallow subsurface flow along the hillslope-riparian-stream continuum. NDVI is widely used as an indicator of vegetation density including leaf area and biomass (Tucker, 1979), and is useful to remove noises by illumination differences and topographic variation in mountainous terrain (Huete et al., 2002). We have applied this method at the regional scale across thousands of headwater catchments to infer changes in runoff generation where it would otherwise be difficult due to sparsely located streamflow gauges.
We investigated how lateral connectivity of forested headwater catchments has changed in response to climate and how this relates to flow dynamics in regional watersheds across the humid, temperate Southern Appalachian Mountains. While this region receives abundant precipitation and is generally considered energy limited, there is evidence that climate change is increasing periods of plant water stress and therefore changing catchment-scale hydrologic partitioning (Hwang et al., 2020). In this study we aimed to (1) Quantify trends in forested headwater lateral hydrologic connectivity using Landsat remotely sensed vegetation patterns from 1984 - 2021, (2) Identify key predictors, including climate and watershed characteristics, that were most explanatory of these patterns, and (3) Relate these changes to regional trends in fluxes and flow dynamics of reference watersheds across the Blue Ridge ecoregion of the southeastern US.
Section snippets
Study region
The Blue Ridge ecoregion is located in the Southern Appalachian Mountains and is approximately 40,940 km2 (US Environmental Protection Agency, 2013). The climate is characterized as marine, humid temperate with abundant precipitation (IQR: 1370–1700 mm y−1), steep slopes (IQR: 20.3–26.6%), and elevation ranging from 200 to 2035 m (Fig. 1). Growing season climate in the region is variable and determined by topography, with higher precipitation and lower temperatures generally found at higher
Headwater lateral hydrologic connectivity analysis
Mean NDVIDown:Up values were centered around 1.0 with a 25th percentile of 0.98, median of 0.99, and 75th percentile of 1.01, indicating denser vegetation upslope in 64% of catchments (Fig. 3). Trends in NDVIDown:Up were significantly negative in 49% of catchments, significantly positive in 10% of catchments, and did not significantly change in the remaining catchments (42%) (Fig. 4). Similarly, trends in NDVISD were significantly decreasing in 36% of catchments, significantly increasing in 12%
General discussion
Across a 40,940 km2 humid, temperate forest landscape, we show that NDVI ratios of upslope to downslope decreased in almost half of the 30,044 continuously forested headwater catchments over the past four decades, primarily due to increasing upslope NDVI (Fig. 3). Increasing vegetation density upslope relative to downslope indicates a decrease in underlying lateral hydrologic connectivity due to increasing upslope ET that reduces shallow subsurface flow, ultimately leaving less water available
Conclusion
We identified widespread homogenization of vegetation density along hydrologic flow paths at the headwater catchment scale in the southern Appalachians. This indicates declines in underlying lateral hydrologic connectivity and shifts in hydrologic partitioning towards more ET and less runoff generation. This also indicates a shift towards more climate and drought sensitive flow regimes across the study site. These trends are likely more closely related to minimum temperature regime rather than
Data availability
All data used in this study was publicly available. Landsat NDVI and land cover data downloaded using Google Earth Engine. USGS streamflow downloaded from (https://waterdata.usgs.gov/nwis/sw). Daymet climate data downloaded from (https://daymet.ornl.gov/). NHDPlusHR headwater dataset downloaded from (https://www.usgs.gov/national-hydrography/nhdplus-high-resolution) and the GAGES-II reference watershed dataset downloaded from (https://pubs.er.usgs.gov/publication/70,046,617). Blue Ridge
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.
Funding
This work was supported by the Center for Geospatial Analytics at NC State University and a Nature Conservancy NatureNet fellowship awarded to Katie McQuillan.
Acknowledgements
This work was supported by the Center for Geospatial Analytics at NC State University and a Nature Conservancy NatureNet fellowship awarded to K. McQuillan. This research was also supported by the Brain Pool program from the National Research Foundation of Korea (NRF) (Grant No.: 2021H1D3A2A01042250) to T. Hwang.
References (91)
- et al.
Cloud detection algorithm comparison and validation for operational landsat data products
Remote Sens. Environ.
(2017) - et al.
A comparison of sap flux-based evapotranspiration estimates with catchment-scale water balance
Agric. For. Meteorol.
(2007) - et al.
Climate, forest growing season, and evapotranspiration changes in the central appalachian mountains, USA
Sci. Total Environ.
(2019) Wide dynamic range vegetation index for remote quantification of biophysical characteristics of vegetation
J. Plant Physiol.
(2004)- et al.
Connectivity at the hillslope scale: identifying interactions between storm size, bedrock permeability, slope angle and soil depth
J. Hydrol. (Amst.)
(2009) - et al.
Overview of the radiometric and biophysical performance of the MODIS vegetation indices
Remote Sens. Environ.
(2002) - et al.
Estimating basin-wide hydraulic parameters of a semi-arid mountainous watershed by recession-flow analysis
J Hydrol (Amst)
(2003) - et al.
Modification of surface energy balance during springtime: the relative importance of biophysical and meteorological changes
Agric. For. Meteorol.
(2020) - et al.
An increasing trend in the ratio of transpiration to total terrestrial evapotranspiration in china from 1982 to 2015 caused by greening and warming
Agric. For. Meteorol.
(2019) - et al.
An improved methodology for predicting the daily hydrologic response of ungauged catchments
Environ. Modell. Software
(1998)