Extended growing seasons and decreases in hydrologic connectivity indicate increasing water stress in humid, temperate forests

https://doi.org/10.1016/j.agrformet.2023.109525Get rights and content

Highlights

  • Assessed trends and drivers of forested headwater hydrologic connectivity.

  • Remotely sensed vegetation patterns used as proxy for connectivity.

  • Connectivity decreased in 49% of catchments, indicating increasing ET.

  • Increasing Tmin (longer growing seasons) associated with decreasing connectivity.

  • Vegetation dynamics important control on water stress under changing climate.

Abstract

Forested headwater catchments are important sources of stable and abundant freshwater resources. Interactions between vegetation and topography influence lateral hydrologic connectivity by altering shallow subsurface flow paths. This in turn influences vegetation density along those paths, and subsequent hydrologic partitioning between localized water use and subsurface flows at catchment scales. Climate change impacts on forests, and the degree to which they reshape feedbacks between evapotranspiration (ET) and hydrologic connectivity, remain unclear. To clarify the extent and drivers of changing lateral hydrologic connectivity, we assessed relative changes in upslope to downslope vegetation density using the Normalized Difference Vegetation Index (NDVI) from 1984 – 2021 in 30,044 forested catchments across the Southern Appalachian Mountains. Increasing upslope NDVI relative to downslope NDVI was used as a proxy for decreasing lateral hydrologic connectivity. We then related changes in connectivity to climate and streamflow dynamics across 28 sub-regional reference watersheds. We found decreases in the ratio of downslope to upslope NDVI in almost half of the catchments (48.5%), primarily due to increasing upslope NDVI. This indicates increasing ET upslope and a decline in lateral hydrologic subsidy to downslope given precipitation. This was also supported by faster streamflow recession and increasing ET estimates relative to precipitation in over half of reference watersheds. The strongest predictor of decreasing connectivity was growing season minimum temperature (Tmin), which increased in 88% of catchments (Mean R2 = 0.27 +/- 0.13). While Tmin is not a dominant atmospheric driver of ET, this pattern has been closely linked to lengthened growing seasons. This suggests that alteration of lateral hydrologic connectivity is mainly driven by ecophysiological responses to changing climate rather than directly by atmospheric drivers. Our results emphasize the importance of vegetation dynamics shifting hydrologic partitioning and driving water limitations even 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)

  • D.E. Rupp et al.

    Information, artifacts, and noise in dQ/dt−Q recession analysis

    Adv. Water Resour.

    (2006)
  • H.J. Tromp-van Meerveld et al.

    On the interrelations between topography, soil depth, soil moisture, transpiration rates and species distribution at the hillslope scale

    Adv. Water Resour.

    (2006)
  • C.J. Tucker

    Red and photographic infrared linear combinations for monitoring vegetation

    Remote Sens. Environ.

    (1979)
  • J.M. Vose et al.

    Hemlock woolly adelgid in the southern appalachians: control strategies, ecological impacts, and potential management responses

    For. Ecol. Manage.

    (2013)
  • A. Weiss et al.

    Partitioning solar radiation into direct and diffuse, visible and near-infrared components

    Agric. For. Meteorol.

    (1985)
  • K.B. Wilson et al.

    Seasonal and interannual variability of energy fluxes over a broadleaved temperate deciduous forest in north america

    Agric. For. Meteorol.

    (2000)
  • G.A. Ali et al.

    Shopping for hydrologically representative connectivity metrics in a humid temperate forested catchment

    Water Resour. Res.

    (2010)
  • B. Beckage et al.

    A rapid upward shift of a forest ecotone during 40 years of warming in the green mountains of vermont

    Proc. Natl Acad. Sci.

    (2008)
  • W. Biggs

    Radiation measurement

  • W. Brutsaert et al.

    Regionalized drought flow hydrographs from a mature glaciated plateau

    Water Resour. Res.

    (1977)
  • P.V. Caldwell et al.

    Declining water yield from forested mountain watersheds in response to climate change and forest mesophication

    Glob Chang. Biol.

    (2016)
  • W. Chen et al.

    Temporal and spatial monitoring of post-fire forest dynamics using time-series

    MODIS Data

    (2014)
  • I. Chuine

    Why does phenology drive species distribution?

    Philosoph. Trans. R. Soc. B

    (2010)
  • F.P. Day et al.

    Forest communities and patterns

  • J.M. Detty et al.

    Topographic controls on shallow groundwater dynamics: implications of hydrologic connectivity between hillslopes and riparian zones in a till mantled catchment

    Hydrol. Process

    (2010)
  • K. Duan et al.

    Implications of upstream flow availability for watershed surface water supply across the conterminous united states

    JAWRA J. Am. Water Res. Assoc.

    (2018)
  • K.J. Elliott et al.

    Forest tree growth response to hydroclimate variability in the southern appalachians

    Glob Chang. Biol.

    (2015)
  • K.J. Elliott et al.

    Long-term changes in forest composition and diversity following early logging (1919–1923) and the decline of american chestnut (castanea dentata)

    Plant Ecol.

    (2008)
  • J.A. Falcone

    GAGES-II: Geospatial attributes of Gages For Evaluating Streamflow

    (2011)
  • Y. Fan

    Groundwater in the earth's critical zone: relevance to large-scale patterns and processes

    Water Resour. Res.

    (2015)
  • C.R. Ford et al.

    Quantifying structural and physiological controls on variation in canopy transpiration among planted pine and hardwood species in the southern appalachians

    Ecohydrology

    (2011)
  • D.C. Frank et al.

    Water-use efficiency and transpiration across european forests during the anthropocene

    Nat. Clim. Chang.

    (2015)
  • K.E. Gleason et al.

    Competition amplifies drought stress in forests across broad climatic and compositional gradients

    Ecosphere

    (2017)
  • M.L. Goulden et al.

    California forest die-off linked to multi-year deep soil drying in 2012–2015 drought

    Nat. Geosci.

    (2019)
  • Gustard, A., and Demote, S. (2009). Manual on low-flow estimation and prediction. Operational hydrology report no. 50,...
  • T.C. Hales et al.

    Topographic and ecologic controls on root reinforcement

    J. Geophys. Res.

    (2009)
  • C.J. Harman et al.

    Power law catchment-scale recessions arising from heterogeneous linear small-scale dynamics

    Water Resour. Res.

    (2009)
  • S. Hawthorne et al.

    Topography may mitigate drought effects on vegetation along a hillslope gradient

    Ecohydrology

    (2018)
  • Hewlett, J.D., and Hibbert, A.R. (1967). Factors affecting the response of small watersheds to precipitation in humid...
  • T. Hwang et al.

    Divergent phenological response to hydroclimate variability in forested mountain watersheds

    Glob. Chang. Biol.

    (2014)
  • T. Hwang et al.

    Simulating vegetation controls on hurricane-induced shallow landslides with a distributed ecohydrological model

    J. Geophys. Res.

    (2015)
  • T. Hwang et al.

    Climate change may increase the drought stress of mesophytic trees downslope with ongoing forest mesophication under a history of fire suppression

    Front. Forests Global Change

    (2020)
  • T. Hwang et al.

    Ecosystem processes at the watershed scale: hydrologic vegetation gradient as an indicator for lateral hydrologic connectivity of headwater catchments

    Water Resour. Res.

    (2012)
  • T. Hwang et al.

    Ecosystem processes at the watershed scale: extending optimality theory from plot to catchment

    Water Resour. Res.

    (2009)
  • T. Hwang et al.

    Nonstationary hydrologic behavior in forested watersheds is mediated by climate-induced changes in growing season length and subsequent vegetation growth

    Water Resour. Res.

    (2018)
  • View full text