Description
Quantifying catchment-scale water cycling over longer periods is
important as wet and dry precipitation cycles can adversely affect blue
(groundwater and discharge) and green (evapotranspiration) water fluxes
and storage dynamics (Orth & Destouni, 2018; Yang, Tetzlaff, Soulsby,
Smith, & Borchardt, 2021). Spatially distributed modelling approaches
are, in many circumstances, essential for spatio-temporal evaluation of
the non-stationarity of flow paths and storages within catchments
(Fatichi et al., 2016). When linked with tracers and water age tracking
such models can also aid in understanding the ecohydrological separation
of water sources (McGuire & McDonnell, 2015). Here, we present a
visualisation of results from a tracer-aided ecohydrological model to
explore spatiotemporal dynamics of water flux-storage-tracer-age
interactions through wet and extreme dry (including the European 2018
drought) cycles. This visualisation extends beyond the traditional
“snapshot” of spatial catchment conditions or temporal depiction of a
single “point”, providing further novel understanding of spatial
differences in catchment response and interactions under wet and dry
conditions.
EcH2O-iso is a tracer-aided ecohydrological model
coupling vegetation-soil-atmosphere energy, water, and tracer
mass-balance with a simultaneous solution of vegetation dynamics and
water age (Kuppel, Tetzlaff, Maneta, & Soulsby, 2018; Maneta &
Silverman, 2013). Energy and water balance are estimated with a top-down
approach from the canopy to the sub-surface. The energy balance is
solved iteratively for temperature (canopy and surface) to estimate
latent, sensible, and ground heat, soil temperatures, and net radiation.
Water balance is solved in five model storages: canopy, surface, and
soil (three storage layers) with vertical movement through all storages,
and lateral movement (kinematic wave) in surface and deep soil water.
Evapotranspiration components are derived from the energy balance. Water
ages and isotopic tracers are estimated in each storage with complete
mixing. The Demnitzer Millcreek catchment (DMC, 66km2)
is a mixed-landuse mesoscale catchment in north-east Germany,
representative of other catchments in the North European Plain. The DMC
has annual mean precipitation and discharge of 560 and 60 mm,
respectively with stream flow being groundwater-dominated (Smith,
Tetzlaff, Gelbrecht, Kleine, & Soulsby, 2020; Smith, Tetzlaff, Kleine,
Maneta, & Soulsby, 2021). Four major landuse units are distinguishable:
agricultural, coniferous forest, wetlands, and broadleaf forest. The
soils are primarily sandy brown earths, with peaty podzolic soils in
wetlands and gley beneath the channels. Discharge and stream chemistry
time-series (1990 – present) and soil moisture, sap flow, soil and
discharge isotopes (2018-present) have been monitored in the main
channel and primary landuses (Gelbrecht, Lengsfeld, Pöthig, & Opitz,
2005; Smith et al., 2021). EcH2O-iso was set up on 250m
square grids for the DMC and simulated on daily time-steps from
2007-2019 using the first two years as spin-up. Multi-criteria
calibration utilized available data to constrain soil storage,
ecohydrological fluxes, and isotopic tracers in the primary landuses
(nine soil and six vegetation parameters for each landuse, Smith et al.,
2021).
The visualisation (Movie 1) shows the bi-weekly spatial and average
changes in catchment storage (in shallow soil and groundwater),
transpiration, specific discharge, transpiration and discharge age, and
tracer dynamics (shallow soil and discharge) over an extended period
from 2009-2019 to provide novel insights of spatially disaggregated
catchment interactions in wet and dry periods. The bottom left panel
shows the profile of catchment landuse for context, with agriculture
dominant in the north, wetlands in the central catchment and forests in
the south. To aid in spatial visualisation, the scales are set to the
5th and 95th percentiles of monthly
averages, with a log-scale for the transpiration ages. Low precipitation
toward the end of the study period caused a long-term decrease in
groundwater (meters of water above field capacity) and summer soil water
storage and an increase in transpiration age.
Relationships of catchment wetness to storage, fluxes, tracers, and ages
were temporally consistent; however, the strength of these relationships
changes seasonally. Higher wetness conditions during the early and mid
growing season (0:07 & 0:21) decreased total catchment transpiration
relative to drier years (0:39 & 0:43) due to lower atmospheric vapour
pressure deficit in wet conditions. Wetter conditions additionally
increased soil water fractionation (higher δ2H) during
the early and mid-growing season due to increased soil evaporation. Late
growing season wetness conditions did not strongly affect transpiration;
however, wetter conditions in winter and higher rainfall years increased
discharge and decrease water ages of transpiration and discharge (0:09
v. 0:57).
Spatio-temporal differences between wet and dry growing season months
are primarily distinguishable by landuse. In particular, wetlands
consistently had higher moisture and lower transpiration, while spatial
differences between agricultural areas in the north and conifer forests
in the south deviated with wetness conditions. Wet conditions in the
early growing season reduced spatial differences of soil moisture (e.g.
0:26 v. 1:01), but resulted in larger spatial differences in stream
water age (lower water age through the conifer areas), and decreased
spatial differences in transpiration water age between conifers and
agricultural areas. Spatial differences in soil moisture were not as
prominent during the mid-growing season regardless of wetness
conditions. However, wetter conditions in mid-summer drove greater
spatial differences in transpiration and channel fractionation, and
lower differences in transpiration age between conifers and agricultural
landuse (e.g. 0:50 v. 0:38). Under these conditions, transpiration rates
were higher in conifer forests, and more water available in shallow
soils for transpiration spatially unified water ages throughout the
catchment. Drier conditions resulted in lower transpiration in conifer
forests compared to agricultural areas. Spatio-temporal variations
during the late growing season showed further consistent trends with
wetness, with greater spatial differences in transpiration and smaller
spatial effects of channel fractionation under wetter conditions (0:15
v. 0:45).
Here, we explored the visualisation of spatio-temporal dynamics
catchment fluxes, storages, tracers, and water ages through wet and dry
periods in a mesoscale, mixed-landuse catchment in Germany. Spatial
variation in transpiration, evaporative fractionation, water age and
soil moisture are directly related to inter-annual variations of
catchment wetness, with additional intra-annual spatial variability
within the growing season. These visualisations of storage-flux-age
dynamics aid in the understanding of the sensitivity of key catchment
regions and could improve knowledge of hydrological catchment
functioning in wet and dry conditions under long-term change. As such,
they are useful tools to communicate the output of complex
ecohydrological models to land managers.