Figure 10. Monthly precipitation from station observations,
precipitation biases from the default (DEF) and wetland scheme (WS)
simulations, and the precipitation difference between WS and DEF
simulations in the summer (May-August) for three-year simulations.
4 Discussion
LSMs and coupled ESMs, reasonable representations of wetland spatial
extents and dynamic water storage are challenging in light of data
scarcity, coarse model resolution, and insufficient understanding of the
physical processes (Ringeval et al., 2012). However, because wetland
extents play a key role in land-atmosphere interactions and carbon
feedback to the climate system, researchers have long been interested in
estimating wetland extents in hydrology-climate simulations from global
to regional scales. For example, the WETCHIMP project gathered 10
participating GCMs for simulating global wetland extents and their
CH4 emissions (Wania et al., 2013; Melton et al., 2013).
Many of these GCMs used prescribed wetland maps from global surveys or
remote sensing products, such as the Global Lake and Wetland Database
(Lehner and D\(\ddot{o}\)ll, 2004) and GIEMS (Prigent et al., 2007), or
used the TOPMODEL-based \(F_{\text{sat}}\) to simulate a subgrid
“saturated” fraction to represent wetlands extents.
Although the TOPMODEL method can simulate some spatial heterogeneity and
temporal dynamics of wetland extent, it generally underestimates both
the maximum value and the seasonal variability. As we showed in Section
2.1, the TOPMODEL-based method in Noah-MP simulates a much lower\(F_{\text{sat}}\) value than the highly dynamic GIEMS product. Here we
provide two possible reasons for the discrepancy between TOPMODEL\(F_{\text{sat}}\) and surface water dynamics from satellites. (1) The
first underlying assumption of the TOPMODEL method requires “steady
state” precipitation and soil moisture heterogeneity, which is more
likely in wet, relatively shallow soils on moderate slopes (Beven and
Kirkby, 1979; Kirkby et al., 2021). However, this is not the case in the
Prairie Pothole Region, where the climate is usually semi-arid and the
large-scale topography is flat with small-scale variation. (2) Another
possible reason for this discrepancy is that the TOPMODEL method
calculates a critical topographic index value when the local water table
is at the surface; this value is used to determine the\(F_{\text{sat}}\) fraction through the integration of its probability
distribution function. However, in the PPR, frozen soils in wintertime
prevent interaction between the soil moisture and groundwater (Ireson et
al., 2013). Therefore, in the TOPMODEL method, the exponential function
will simulate less seasonal variation in the surface water dynamics. A
large portion of global wetlands and peatlands are located in high
latitude regions where winter soil freezing is very common.
In our modification of the \(F_{\text{sat}}\) formulation, we used the
first layer of soil saturation to indicate the sub-grid spatial extent
of the saturated portion – the extent of wetlands. This method
empirically assumes the grid cell mean soil moisture saturation can be
translated into a spatial fraction for surface saturation and shows a
highly variable \(F_{\text{sat}}\) value compared to the default
TOPMODEL method, in terms of the maximum and minimum extent, and
seasonality (Section 3.1). Moreover, we also incorporate a spatially
varied maximum \(F_{\text{satmx}}\) map from the GIEMS product to
replace the default global mean value (0.38) in Noah-MP and WRF. Both
these modifications improve the spatial heterogeneity and the temporal
dynamics of wetland extents in the PPR.
Additionally, we incorporated a dynamic surface water storage scheme to
simulate the hydrological processes in wetlands. Although this scheme is
simple, we aim to capture three important processes – the filling of
wetlands by snowmelt and rainfall, the evaporation of wetland water into
the atmosphere, and the excess water spilling to surface runoff. These
three processes are the key components in the wetland water and energy
cycle during the warm season open-water period. Our results showed
increase of ET with a decrease of surface runoff and an increase of
latent heat with decreases of sensible heat. This finding aligns with
our expectations, as well as with previous VIC model wetland and lake
simulations in the U.S. Midwest region (Mishra and Cherkauer et al.,
2010).
Moreover, our scheme provides greater potential to explore wetlands’
feedback to the atmosphere in coupled WRF-NoahMP-Wetland simulation. In
the default simulation, which already includes the MMF groundwater
scheme (Barlage et al., 2015, 2021), warm biases still exist at about
1~3 degrees in the U.S. Great Plains. Without the
groundwater scheme, the summertime warm biases could be as high as
4~6 degrees. By adding the wetland scheme on top of the
MMF groundwater scheme, the warm biases in the U.S. can be further
reduced by 0.5~1.5 degrees, but it also introduces
1-degree cool biases in Southern Manitoba, where wetland extents are
large. While the temperature cooling effect is evident, wetland feedback
to precipitation is less clear and is more ambiguous. A previous study
using WRF with a prescribed soil moisture threshold to indicate wetlands
in the Great Plains at coarser resolution (12-km) also showed a
temperature cooling effect, but the precipitation effect was negligible
(Capehart et al., 2012).
One of the highlights of this study is the wetland cooling effect to the
atmospheric temperature. Previous studies have documented this effect in
detail, but they have been specific to different wetland characteristics
and dominant vegetation types (Pitman, 1991; Bonan, 1995). In our study,
we used general open-water storage to characterize wetland interactions
with the atmosphere, omitting these variations in specific wetland types
but gaining more generic conclusions in a much larger region. The
wetland cooling effect on temperature, especially during extreme
heatwave events, echoes a previous study in the Central U.S. where we
found land surface characteristics could effectively reduce the
frequency, intensity, and duration of extreme heatwaves (Zhang et al.,
2018). However, more pronounced cooling occur in non-heatwave years
(2005 and 2007) than in 2006, because the cooling effect also depends on
water availability, hence, cannot be too dry.
In recent years, the tradeoffs between agriculture and wetland
conservation has been a serious topic of discussion among the public,
universities, and government agencies. It has been shown that the
agricultural land expansion at the cost of wetland drainage increases
the risk of emerging flooding in springtime (Dumanski et al., 2015;
Pattison-Williams et al., 2018). Wetland drainage also results in
increased nutrient export (Badiou et al., 2018; Wilson et al., 2019) and
carbon release to the atmosphere (Badiou et al., 2011). This study
suggests that the loss of wetlands for croplands also reduces resilience
to drought and high temperature, which may cause crop failures due to
water and heat stress (Hatifield, 2016).
However, the loss of wetlands to agricultural, industrial and
residential land is not confined to the PPR but are common problems
worldwide and require humans’ attention (The Rasmar Convention 2007;
Nature Geoscience, 2021). These land use modifications not only threaten
the local environment but also contribute to the global carbon balance
and eventually cause problems for human beings. Understanding the
effects of development is challenging. It is hoped that these threats to
the future can inspire future studies on wetlands for their
hydrological, climatic, ecological, environmental functions and that
solutions can be found for humans to interact with nature peacefully and
sustainably.