Leveraging SMAP to improve understanding

of carbon-water-climate interactions

Allison Steiner (PI)

Gretchen Keppel-Aleks (co-I)
Roger De Roo (co-I)
Department of Atmospheric, Oceanic and Space Sciences
University of Michigan, Ann Arbor, MI

A proposal to the NASA ROSES A.22

Science Utilization of the Soil Moisture Active-Passive (SMAP) Mission (Amended)

TABLE OF CONTENTS

Scientific Narrative. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.0 Project Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.0 Relevance to NASA and SMAP Priorities. . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3.0 Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
4.0 Methodology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5.0 Work Plan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
5.1 Task 1: Evaluation of modeled soil moisture state. . . . . . . . . . . . . . . . . 9
5.2 Task 2: Process level water-energy-carbon metrics . . . . . . . . . . . . . . . 10
5.3 Task 3: Parameter testing.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
5.4 Task 4: Understanding long-term impacts. . . . . . . . . . . . . . . . . . . . . . 14
6.0 Project Team, Management and Timeline. . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
6.1 Project team and qualifications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
6.2 Project management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
6.3 Timeline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Biographical Sketches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
A.L. Steiner. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
G. Keppel-Aleks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
R. De Roo .................................................... 24
Summary of Work Effort. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Current and Pending Support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
A.L. Steiner. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
G. Keppel-Aleks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
R. De Roo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Budget Justification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Facilities and Equipment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Detailed Budget. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

Leveraging SMAP for improved understanding

of carbon-water-climate interactions

1.0 Project Overview

Soil moisture is a key driver of carbon and water cycles in terrestrial regions, and ecological
functioning depends on water availability. However, soil moisture remains one of the most
poorly constrained components of land-atmosphere models, and comprehensive datasets are
needed to improve our understanding of biosphere-atmosphere exchange of water and carbon.
New observations from the SMAP satellite launched in 2015 present an opportunity to
understand the spatial and temporal heterogeneity in soil moisture not captured by point-based
measurements. In this proposal, we will use SMAP radiometer data to probe the coupling of
water and carbon as limited by soil moisture in the global climate system (Figure 1).
We will use a global climate model (CESM-CLM4.5) with SMAP satellite-derived Level 2
radiometer and Level 4 model data to investigate the following scientific questions focused the
temperate-to-boreal forested ecosystems in North America:
1. How does modeled soil moisture compare spatially and temporally with SMAP observations?
2. How do spatial and seasonal variations in soil moisture control evapotranspiration and carbon
fluxes?
3. What are the key model parameters that influence soil moisture state and water-carbon
feedbacks?
While we will use a single coupled
model to quantify the relationships
between soil moisture, water and
carbon fluxes, we will go beyond
model evaluation with parameter
testing to identify key variables
necessary to understand these coupled
processes. The evapotranspiration and
photosynthesis parameters can inform
a broad range of other land surface and
climate models via improved
understanding of coupled water and
carbon fluxes. By looking at the first
seasonal cycles of SMAP retrievals
and testing observed relationships with
the model, we will elucidate the role of
soil moisture on coupled carbon and
water fluxes in the temperate to boreal
transition zone of the Northern
Hemisphere. These short-time scales
can inform broader climate-timescale
studies of water and carbon.

2.0 Relevance to SMAP and NASA Priorities

The proposed work will address the scientific priority identified in the program call “to enable
advances in the study of the water, carbon and energy cycles, especially on those topics that

2
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

deal with the intersections of these cycles.” Our collaborative research will use SMAP
radiometer data to improve our understanding of the coupled water-carbon-climate system.
SMAP data will facilitate process-based studies with the CLM, a land surface model used in
weather and climate models as well as data assimilation systems. Specifically, this work will
address SMAP specific goals to “understand processes that link the terrestrial water, energy and
carbon cycles” and “quantify net carbon flux in boreal landscapes.” (SMAP Handbook Project
Goals 1 and 3).

3.0 Background
3.1 Soil moisture-water-carbon relationships
Soil moisture acts as a key driver in the climate system by controlling water and carbon
exchange between the surface and the atmosphere [Seneviratne et al., 2010]. In vegetated
regions, this relationship becomes more complex because vegetation relies on water from its
rooting system. As a result, the amount of soil moisture present controls the photosynthetic and
evaporative demand of vegetated regions. While other environmental drivers such as
temperature and precipitation are often used to explain the exchange of water and carbon, the
role of soil moisture and its limiting effects on ecosystem functioning is possibly more
fundamental but still very poorly constrained. Therefore, understanding the soil moisture-water-
carbon interactions is key to understanding regional hydroclimatology and precipitation
[Dirmeyer et al., 2009; Seneviratne et al., 2010], as well as understanding the regional and global
projections of the terrestrial carbon [Suyker et al., 2003].
Terrestrial vegetation mobilizes moisture from the subsurface to the non-woody parts of
the plant (e.g., stems and leaves) through its rooting system. Green biomass converts radiant
energy to chemical energy through photosynthesis, the process that converts CO2 into
carbohydrates and new biomass. To draw CO2 from the atmosphere into the leaf for fixation,
leaves have openings on their surface known as stomates. When the stomates open to allow CO2
into the leaf, water vapor inexorably escapes due to the strong gradient from the nearly saturated
environment inside the leaf to the relatively dry atmosphere. This biological process of stomatal
control inherently couples the water and carbon cycle when vegetation is present. In this regard,
it is impossible to examine carbon exchange without understanding the relative role of
evapotranspiration and the limitations of soil moisture.
In vegetated regions, evapotranspiration (ET) is the sum of evaporation from the ground,
evaporation from water stored within the canopy, and transpiration or the release of water from
the internal plant tissues during stomatal opening. Ground-based observations from flux towers
[Baldocchi et al., 2001] and satellite-derived ET [Mu et al., 2011] both estimate total
evapotranspiration from ecosystems. However, models still show difficulty in accurately
simulating ET in vegetated regions over a range of time scales from daily [Matheny et al., 2014]
to interannual [Jung et al., 2010]. To understand the role of soil moisture on modeled ET,
observationally based metrics indicate that ET is still underestimated by models, particularly
during the summer months when soil moisture may not be properly represented. For example,
NLDAS experiments show that even with data assimilation techniques, models still
underestimate summer ET [Xia et al., 2015]. At the climate time scale, [Bryan et al., 2015]
shows that the Community Land Model (CLM) under predicts the observed temperature-
evapotranspiration relationship in mid-latitude deciduous forests, which may be related to
erroneous soil water content or poorly parameterized transpiration processes.

3
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

Because of the tight coupling of transpiration with carbon, soil moisture also has the
potential to influence global gross primary productivity (GPP, or the net carbon assimilated
during photosynthesis after accounting for autotrophic respiration) on a range of time scales (e.g.
[Lei et al., 2014] ). Global scale simulations suggest that moisture stress in temperate Northern
Hemisphere ecosystems is an important driver of interannual variability of carbon sinks [Keppel-
Aleks et al., 2014], and the precipitation-productivity relationship is largely derived through
vegetation uptake of soil water. Until recently, the terrestrial carbon sink has been thought to
respond mainly to tropical temperature variability [Wang et al., 2013]. However, when
variations in the
atmospheric CO2
growth rate are
analyzed as a
function of latitude,
variations in the
growth rate can be
traced to several
environmental
drivers, including
drought (Figure 2).
Most relevant for
the proposed
research,
temperature and
drought anomalies
(measured by
Palmer Drought
Severity Index) in the Northern Hemisphere mid- to high-latitudes contribute 0.23 Pg C y-1
variability to the global carbon sink (or about 10% of the average annual sink). We note that
temperature and drought stress co-vary, and that means soil moisture may affect the temperature
anomaly over land via the key role of latent heat in determining the surface energy balance. The
process-based research proposed here will strengthen our understanding of the mechanisms that
link water-carbon cycling, regional differences and temporal trends.
While most studies suggest that soil moisture is a strong driver in water and carbon
exchange, the lack of broad scale soil moisture observations has precluded clear observable
metrics to quantify soil moisture’s role. Frequently, the role of soil moisture as a driver has been
recently explained in the energy-limited vs. moisture-limited framework (e.g., as reviewed in
[Seneviratne et al., 2010]), yet clearly defining the soil moisture limitations is still needed to
understand the role of soil moisture limitations on water and carbon fluxes (Science Question 2).

3.2 Regional Focus: Temperate to Boreal Transition

Forested areas are difficult regions to understand the role of soil moisture, yet they
represent one of the most important ecosystem types to understand. Challenges in understanding
soil moisture in these regions include (1) the density of vegetation, which complicates remote-
sensing techniques retrievals, (2) the complexity of the root structure, with different types of
vegetation contributing different root zones, and (3) the heterogeneity of the rooting systems and

4
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

subsequent soil moisture. Despite these challenges, quantifying the role of soil moisture on
forest ecosystem functioning is greatly needed.
In this proposal, we will focus on the temperate to boreal transition region over North
America (Figure 3). Boreal forests represent the regions across Eurasia and Northern America
just south of the tundra line, and they can be either evergreen needleleaf forest (ENF) or in some
cases, deciduous broadleaf forest (DBF). In the temperate forests of North America, trees are
dominantly DBF allowing several months of the year without foliage (~November-March).
Modeled soil moisture in the region shows a strong seasonal cycle, with a 50% variation in
boreal regions and 30% variability in temperature regions, with substantial spatial heterogeneity
(Figure 3). Overall, we select the transition from temperate to boreal forests to capture a range
of the northern hemisphere carbon sink and understand the role of soil moisture across this
transect.

3.3 Known deficiencies in modeled soil moisture and its impacts

The NASA SMAP mission was initiated to improve the observations of soil moisture at
both spatial and temporal scales. From the perspective of climate and weather, the use of this
data, either as assimilated into models or used for evaluation, will improve our understanding of
soil moisture in the climate system. This remains a crucial science need. For example, the North
American Land Data Assimilation System (NLDAS) still shows that evapotranspiration is
underestimated in the summer and winter by most NLDAS models, particularly for DBF and
ENF [Xia et al., 2015]. Additionally, climate models are currently lacking the broad spatial and
temporal scales needed to fully evaluate soil moisture for its role in climate processes. In this
proposal, we plan to leverage SMAP radiometer data to identify and quantify the role of soil
moisture on temperate to boreal forested ecosystems in the Northern Hemisphere.

4.0 Methodology
As we investigate the role of soil moisture on evapotranspiration and carbon fluxes, we will
use a single land model (Section 4.1) coupled a global climate model (Section 4.2). These

5
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

simulations will be evaluated and tested with Level 2 radiometer and Level 4 model SMAP data
(Section 4.3) and other ground-based observational sets (Section 4.4).

4.1 The Community Land Model (CLM) version 4.5

The CLM4.5 is a land surface model developed for use with the NCAR Community Earth
System Model (CESM) that has gone through five model revisions to arrive at its present
version, CLM4.5 [Oleson et al., 2013]. Here we briefly review three CLM4.5 submodules most
relevant for the proposed work, including soil moisture, evapotranspiration, and carbon.
Soil Moisture: CLM4.5 defines the soil column with fifteen unevenly spaced soil layers (10 with
hydrology) and up to five layers of snow depending on snow depth. For surface hydrology and
determining the input of water to the soil column, CLM4.5 parameterizes surface runoff
(SIMTOP; [Niu and Yang, 2006a]) and surface water storage, and uses these two terms to define
infiltration into the soil column. Within the soil column, the model accounts for sub-surface
runoff and interactions with groundwater (SIMTOP), gradient diffusion, and canopy
transpiration (through defined root zones in each soil layer). The soil water flux is a function of
the hydraulic conductivity (k) and the soil matric potential (ψ) of the soils, and is calculated
using a modified Richards equation [Zeng and Decker, 2009]. The hydraulic soil properties (k
and ψ) are based on the volumetric soil water and the soil texture, weighted as a combination of
mineral (sand, silt, and clay; [Clapp and Hornberger, 1978]) and organic components [Lawrence
and Slater, 2008]. CLM4.5 also explicitly accounts for phase changes in individual soil layers.
Additionally, the model accounts for supercooled soil water (water co-existing with ice before it
freezes) with a freezing point depression modification that is a function of soil texture [Niu and
Yang, 2006b]. Energy needed for soil moisture phase transitions are incorporated into surface
energy balances. The default option for soil moisture initialization is to specify the ten soil
levels with a soil water content of θi = 0.15 mm3 mm-3 for all levels, although all proposed model
simulations will include a ten-year model spinup to allow full soil moisture equilibration. Root
zone structure is determined by the plant functional type (PFT) and uses two PFT-specific
parameters (ra and rb) to define the amount of roots in each layer [Zeng, 2001].
Evapotranspiration (ET): The CLM4.5 evapotranspiration parameterization uses an energy-
budget based approach to calculating ET, which differs from many hydrology-based land models
that use a water-budget approach [Lofgren et al., 2013]. CLM uses Monin-Obukhov similarity
theory to determine flux quantities, which vary depending on the composition of the underlying
surface (e.g., non-vegetated or vegetated) with aerodynamic parameters that modulate exchange.
Evapotranspiration (ET) is computed as a function of the specific humidity gradient between the
soil and/or canopy surfaces (qsfc) and the overlying atmosphere (qa) as a function of the surface
resistance (r):
𝑞! − 𝑞!"#
𝐸𝑇 = −𝜌
𝑟
Over vegetated surfaces, r is a function of the aerodynamic resistance from the surface, the leaf
boundary layer resistance, and the stomatal resistance for both sun and shade canopy leaves. In
this manner, transpiration from vegetation is included in the ET calculation. Temperatures and
fluxes are solved using a Newton-Raphson iteration method.
Carbon: CLM4.5 simulates carbon assimilation using a modified Ball-Berry model to relate leaf
stomatal conductance to net photosynthesis [Collatz et al., 1991]. Net photosynthesis is
calculated differently for C3 plants [Farquhar et al., 1980], which dominate our study region, and
for C4 plants [Collatz et al., 1992]. In both cases, the photosynthesis rate is determined by co-

6
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

limitation among the Rubisco-, light-, and export-limited rates [Bonan et al., 2011]. CLM4.5 has
been revised to better parameterize light use for sunlit and shaded leaves [Bonan et al., 2012],
and the maximum rate of carboxylation, VCmax, is calculated separately for sunlit and shaded
leaves since nitrogen allocation decreases with depth in the canopy. Leaf area index (LAI) in
CLM4.5 is derived from a monthly climatology of multi-year satellite records. Stomatal
conductance of the leaves, and thus carbon assimilation, is multiplied by a soil water stress
factor, βt, which ranges from 1 when soil is wet to nearly zero when soils are dry, depending on
the simulated soil water potential, PFT-dependent root distribution, and PFT-dependent wilting
factor.

4.2 The Community Earth System Model (CESM)

The Community Earth System Model (CESM) integrates the CLM described above into a
fully coupled general circulation model with atmosphere, ocean, land, and sea ice components
[Hurrell et al., 2013]. CESM was developed from model components in the Community Climate
System Model (CCSM, [Gent et al., 2011]), with the addition of a fully prognostic carbon cycle.
CESM can be run with all components coupled, but for the proposed research, we will run CLM
either coupled to the Community Atmospheric Model (CAM; [Neale et al., 2013]) or with an
atmospheric reanalysis product (e.g. CRU-NCEP) and prescribed boundary conditions for ocean
and ice components with a 100km global resolution. CESM can simulate net carbon fluxes using
the biogeochemistry model (BGC), as opposed to only the gross photosynthesis simulated by the
biogeophysical model in CLM (described above). When BGC is active, carbon uptake, allocation,
and release via respiration from terrestrial ecosystems are predicted explicitly [Lindsay et al.,
2014]. Leaf phenology is likewise prognostic, which may have feedbacks on simulated
evapotranspiration rates and thus soil moisture within the model.
Within CLM4.5, there are two configurations available for biogeochemistry: CLM4.5-BGC
and CLM4.5-CN. CLM4.5-CN includes the effects of nitrogen limitation beyond the impact of
nitrogen on VCmax (Section 4.1) but may damp feedbacks between the carbon cycle and other
components of the Earth system [Thornton et al., 2007]. We will use CLM4.5-BGC, as this
version contains improved representation of soil carbon pools, which are now based on the
CENTURY model [Parton et al., 1987] and include better vertical resolution and improved
decomposition rates. These changes have led to improved agreement with observational
benchmarks for gross primary production (GPP) at FLUXNET sites within temperate North
America. Despite these improvements, there are substantial site-to-site differences in the ability
of CESM to capture to magnitude and phasing of carbon fluxes. These changes made little
impact on the phasing and magnitude of ET: both CLM4CN and CML4.5-BGC have peak ET
values that are too high by 20% and that peak in June, one month early. The relationships
between ET and either temperature or precipitation are, however, improved in the new model
relative to observations.

4.3 SMAP data

This proposal will integrate Level 2 radiometer and Level 4 model SMAP data with a global
climate model to quantify the role of soil moisture on terrestrial water and carbon cycles. While
the use of the now defunct active sensor would certainly have been useful, our work will use the
Level 2 radiometer data as the primary data source. Based on the proposed resolution of our
CESM experiments described above (100 km), our scientific questions are not greatly affected
by the change in retrieved resolution (e.g., Level 2 36 km radiometer soil moisture versus the

7
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

Level 2 9 km active/passive combined soil moisture) and therefore we will still be able to
develop a unique application of the Level 2 radiometer products. We propose to use the Level 2
radiometer soil moisture product (L2_SM_P) on the 36 km grid for CESM-CLM model
evaluation and process-level analysis. This product, estimated for the top 5cm of the soil
column, will be compared with the equivalent depth from the CESM-CLM (approximately the
first three soil layers). Additionally, we will use the Level 4 surface and root zone soil moisture
(L4_SM) net ecosystem exchange product (L4_C) that uses the land surface model to compare
Level 4 NEE (assumed to be 36 km resolution due to revised data sources) and root zone soil
moisture with those from the CESM-CLM.
Due to the differences in scale of the SMAP L2 and L4 products and the global climate
model, we will aggregate the finer-resolution SMAP data (36 km) to the CESM-CLM climate
model grid cell resolution (100 km). Within each CESM-CLM grid cell, we will compare the
PFT and land cover type variability in conjunction with the spatial soil moisture variability to
understand the role of the SMAP spatial resolution upscaling.

4.4 Other observational datasets

To provide an observational constraint on the water and carbon fluxes, we will use a suite
of point-based measurements from the FLUXNET network [Baldocchi et al., 2001]. The
FLUXNET network includes the sites in both the US and Canada, and provides eddy-covariance
measurements of ecosystem-level exchange of water and carbon. Sites for this proposed study
were selected based on (1) presence in the temperate-to-boreal transition (Section 3.2; Figure 3)
and including DBF, ENF or mixed forest sites, (2) currently active sites to use with the most
recent SMAP data, and (3) the availability of site-based soil moisture data. Selected sites are
listed in Table 1. We also note that the FLUXNET footprints are relatively small (~1km2 with
site-dependence), which is much finer than the SMAP L2 resolution (36 km).
While there are often soil moisture sensors at FLUXNET sites, their quality is uncertain with
distinct site-to-site data availability. We will evaluate these datasets when available, and also
include ground-based observations from the USDA Soil Climate Analysis Network (SCAN,
http://www.wcc.nrcs.usda.gov/scan/). These sites have hourly soil moisture at depths of 5, 10,
20, 50 and 100 cm during the present-day period. Because our first evaluation will be for the
year 2015, we anticipate that ground-based data from this network will be available for our
process-based analysis (Task 2) scheduled to start in the second half of 2016.

8
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

5.0 Work Plan

5.1 Task 1: Evaluation of modeled soil moisture states versus observations
Task 1 will address Science Question 1 regarding the ability of the CESM-CLM to simulate
soil moisture over the temperate-boreal transition region of North America (Figure 3). We will
use a combination of SMAP and ground-based data to evaluate the modeled soil moisture state
during the first full seasonal cycle of data (e.g., 2015-2016).
Task 1a: VWC constraints in temperate regions One key limitation in the SMAP retrieval is the
vegetation water content (VWC), which can limit retrievals when VWC > 5 kg m-2. ENF have a
mean VWC of 12.59 kg m-2 (ranging from 9.79 to 13.97 for the 5th and 95th percentiles,
respectively) and DBF have a mean VWC of 11.23 kg m-2 (ranging from 9.54 to 12.40 for the 5th
and 95th percentiles, respectively) [Chan et al., 2013]. In theory, this excludes retrievals over
most of the forested regions of the Northern Hemisphere. Despite this limitation forests remain
one of the key regions with an urgent need for understanding soil moisture. Under these
constraints, we plan to use the
time periods when data is
available (approximately
November- March in
Northern Hemisphere
temperate regions) when
leaves are not present in
temperate regions to evaluate
the simulated soil moisture
state (Figure 4). This off-
season data, in conjunction
with the freeze/thaw state of
the soil will allow us to
determine beginning of
season and end of season data
points for use to evaluate the
model. In this manner, we
can assess the CESM-CLM soil moisture state when VWC ≤ 5 kg m-2 in these regions by using
the modeled seasonal cycle to fill in the SMAP growing season data in a manner consistent with
the modeled seasonal cycle and the beginning and end season points. This approach will also be
cross-evaluated with Level 4 soil moisture data (L4_SM), although we note that the L4_SM
product will be based on a single land model that may have different soil moisture states than the
CLM. However, this evaluation will be useful for the L4 products as well by providing an
alternate soil moisture simulation based on a different model. With the first seasonal cycle of
vegetation (e.g., 2016), we can examine the efficacy of this approach and the ability of the model
to simulate soil moisture during the temperate growing season. This technique will not be
acceptable for most ENF that do not experience the broadleaf senescence, but we can use nearby
DBF areas to test the beginning and end season points in the boreal forests.
Task 1b: Evaluation of modeled soil moisture state We will conduct a baseline decadal CESM-
CLM simulation that encompasses the first full annual year of the SMAP retrieved dataset (e.g.,
2015) and also includes a ten-year model spinup to allow soil moisture to full equilibrate. We
will use this baseline simulation to conduct our first evaluation of modeled soil moistures state
by the CESM-CLM as compared to observed metrics, including:

9
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

(1) For surface soil moisture, how does CESM-CLM modeled surface soil moisture compare
with observed SMAP surface (5cm) Level 2 radiometer-derived soil water at sites that meet
the VWC criteria?
(2) For root zone soil moisture and sites with vegetation, how does CLM modeled soil moisture
compare to SMAP Level 4 derived data?
(3) Can we define different soil moisture states by region to evaluate the ability of this specific
land model to capture SMAP Level 2 radiometer and Level 4 data products?
For example, Figure 3 shows the interannual variability produced by the CESM-CLM for
temperate and boreal regions. The boreal region shows a late-summer drying with relatively low
interannual variability, yet the temperate forests do not exhibit as much variability in the summer
drying. We will examine the potential for the CLM to be “locked in” to a soil moisture state for
different regions and components of the seasonal cycle.
Task 1c: Evaluation with ground-based observations The third evaluation task will focus on
understanding observed relationships between soil moisture from SMAP (including both L2 and
L4 products), ground-based observations from the SCAN network, and the modeled CLM soil
moisture. While we will be limited to one full seasonal cycle in Year 1 of the project, we can
evaluate past relationships for the other datasets (SCAN, FLUXNET) versus that simulated to
understand soil moisture transition states (e.g., dry to wet) in other regions. We will focus on the
temperate-boreal transition region shown in Figure 3, and use the ground-based data available
from SCAN and FLUXNET towers (Table 1).

5.2 Task 2: Process-level water-energy-carbon diagnostic metrics

In Task 2, we will move beyond Task 1 evaluation to a process-level understanding of the
interactions between soil moisture and the water-energy-carbon dynamics (Science Question 2).
We focus on aboveground metrics (e.g., GPP, ET) as affected by belowground processes.
Despite the selection of a single model (CESM-CLM), the parameterization of canopy variables
(e.g., stomatal conductance and photosynthetic rates) do not vary substantially between most
Earth system models as they do for subsurface hydrology. For example, most canopy treatments
use a big-leaf approach [Sellers et al., 1996] and base photosynthetic parameterizations on the
Ball-Berry model [Collatz et al., 1991]. We will focus on understanding how vegetation in
climate models responds to subsurface hydrology and soil moisture with model-observation
evaluations.
Task 2a: Evapotranspiration metrics To understand the process-level relationships between ET
and soil moisture, we will evaluate observed and modeled ET drivers versus atmospheric and
surface variables. Our earlier work with a regional climate used the relationship between the
vapor pressure deficit (VPD) and ET to evaluate the ability of the model to capture the observed
metrics model [Bryan et al., 2015]. VPD, defined as the difference in the saturation vapor
pressure and the actual vapor pressure of the atmosphere (es(Ta) – ea), is a useful metric for the
evaporative demand that drives overall ET. For example, the observed ET-VPD relationship at
the UMBS FLUXNET site is not reproduced by a regional climate model (using CLM3.5) in the
summer (JJA; red) and fall (SON; yellow) (Figure 5). Evaluation of VPD versus es(Ta)
highlights site differences that cannot be explained by temperature or leaf area index alone,
suggesting that poorly constrained soil moisture may be limiting ET [Bryan et al., 2015].
Including correlations between ET and soil moisture using SMAP and other ground-based
observations can disentangle these relationships and highlight energy-limited vs. soil moisture

10
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

limited systems in
the models. This
will allow us to use
SMAP observations
to quantify the role
of soil moisture in
the climate system
and improve global
and seasonal
estimates of ET.
Task 2b: Carbon
cycle metrics We
will similarly
analyze the
relationships
between carbon fluxes and environmental drivers at FLUXNET sites. We will use gross primary
production (GPP), which represents the gross carbon uptake via photosynthesis in an ecosystem
and net ecosystem exchange (NEE), which represents the residual carbon uptake between GPP
and respiration. Like GPP, heterotrophic respiration in the soil depends both on temperature and
soil moisture [Saleska et al., 2003; Davidson and Janssens, 2006].
As described above for evapotranspiration, we will regress SMAP observations in the
vicinity of FLUXNET sites against carbon fluxes measured on FLUXNET towers. Although
NEE is the raw quantity measured using the eddy covariance technique, GPP and thus total
respiration can be inferred by considering daytime and nighttime measurements separately. One
limitation of this approach is that for sites within high VWC canopies, we will not be able to
infer a mid-summer soil moisture level from SMAP observations.
We hypothesize that the soil moisture-GPP relationship will be stronger than the soil-
moisture NEE relationship. We will investigate multivariate linear and nonlinear regressions to
account for the combined role of temperature and soil moisture on controlling ecosystem carbon
fluxes. The slopes calculated for each metric (Temperature-GPP, soil moisture-GPP,
temperature-NEE, and soil moisture-NEE) will be an important metric for considering parameter
adjustments within the CLM model, but the scatter between carbon flux data and environmental
driving data will also provide information because we expect that both drivers affect carbon
fluxes in nature.
After establishing metrics at FLUXNET sites, we will leverage solar-induced fluorescence
(SIF) data from the OCO-2 satellite [Frankenberg et al., 2014]. SIF scales with instantaneous
canopy photosynthesis or GPP. Thus, SIF can infer how much carbon ecosystems
photosynthesize while XCO2 can be used to infer how much of the carbon remains in terrestrial
ecosystems at seasonal or longer timescales. SIF is well-suited for the proposed research, since
the SIF signal captures the response of carbon uptake to water stress even in situations where
conventional remote sensing vegetation indices, which essentially measure greenness, remain
stable [Lee et al., 2013]. We anticipate that by coupling these two satellite data products, we can
more clearly illuminate the effect of mean state on seasonal soil moisture impacts on GPP.
Task 2c: Precipitation recycling An additional metric for quantifying the role of soil moisture on
terrestrial feedbacks through ET is with the precipitation recycling term:
!! !
!
= !!!"

11
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

where PE/P is the fraction of precipitation derived from evaporation and IF is the advective
moisture inflow into a specified region [Schar et al., 1999]. This metric can identify the impact
of soil moisture via ET on regional precipitation and climate. For the specified temperate and
boreal regions in Figure 3, we will calculate the modeled recycling ratios. Because the models
have sufficient information to calculate E over broad spatial regions and IF from the atmospheric
output (not possible for observations due to the dependence on atmospheric reanalysis products,
which are known to have biases), we can assess the potential atmospheric feedback errors that
derive from soil moisture. In this manner, we can use the SMAP data to understand the strength
of the soil moisture response in the atmosphere. While there are other more complex metrics to
determine land-atmosphere coupling, this is one that requires no additional, fixed-seasonality soil
moisture as used in other methods (e.g., GLACE [Koster et al., 2004]).

5.3 Task 3: Parameter testing

The third task addresses Science Question 3 by identifying key canopy model parameters
that influence soil moisture state and land-atmosphere feedbacks. Two existing land surface
model evaluation projects can be mined to understand and focus on specific deficiencies in the
soil moisture-water-carbon system. These include:
(1) The International Land Model Benchmarking (ILAMB) project [Mu et al., 2013] using
results from Earth system models participating in the fifth Coupled Model
Intercomparison Project (CMIP5; [Taylor et al., 2012]), and
(2) The North American Land Data Assimilation (NLDAS; [Mitchell et al., 2004]) and
NLDAS2 [Xia et al., 2012] experiments, which are designed to reduce errors in the soil
moisture and energy fluxes for weather prediction.
While the subsurface hydrology is no doubt an extremely important component of the
modeling system, other studies have focused on hydrology-based parameter testing (e.g.,[Hou et
al., 2012]). Here we specifically focus on understanding how soil moisture alters the fluxes of
water and carbon to the atmosphere and to elucidate how vegetation-driven model parameters
are altered by soil moisture state (e.g., Science Question 3). By testing alternative parameters
within CESM against the SMAP dataset and deriving functional relationships among climate,
water, and carbon variables, our results with CESM-CLM can be extended to other modeling
frameworks.
Preliminary analysis conducted as part of the ILAMB study [Mu et al., 2013] show diverse
patterns of coupling between climate, water, and carbon. Models simulate peak summer
evapotranspiration that is too high and too early relative to an observational benchmark.
Moreover, the models also show high ET rates relative to precipitation rates. Of interest for the
our study region of the North American temperate-boreal transition zone, the observational
benchmark shows ET to be nearly flat as a function of mean annual temperature between 0 and
15°C. In contrast, most models show that ET increases between 0 and 15°C, with many models
showing a doubled ET rate within this temperature range.
From the NLDAS experiments and analysis, the use of a suite of land surface models can
provide insights into model deficiencies that may drive the simulation of soil moisture. While
NLDAS focuses on predominantly subsurface processes and does not assess carbon fluxes, the
information on ET can provide some information for parameter testing. It is important to note
that hydrology-based models (e.g., the VIC model used in the NLDAS study) often calculate
potential ET using different parameters than those used in climate models [see discussion above
and in Weiss and Menzel, 2008; Lofgren et al., 2013]. Results from the first two phases of

12
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

NLDAS suggest that the participating models (including CLM in NLDAS-1) generally
underestimate ET, particularly in the summer months [Xia et al., 2012; Xia et al., 2015].
NLDAS-2 yielded greater inter-model agreement that has been attributed to improvements in
forcing data and some model parameterizations for potential ET [Xia et al., 2012]. [Xia et al.,
2015] evaluation of NLDAS-2 shows both seasonal and PFT based deficiencies, e.g., where the
summer and winter show greater deficiencies in modeled ET versus observations, and that DBF
evaluates better than ENF.
These cross-model differences from the ILAMB and NLDAS experiments suggest that
different land model components simulates the coupling between the water and carbon cycles
differently, which is not a surprising fact but one that is not well tested. In all cases, we will
conduct short-term (e.g., 1 year) CESM-CLM simulations that modify individual parameters and
re-evaluate metrics from Task 2. We propose testing three canopy-based factors that can affect
the simulation of soil moisture in global climate models, including:
Task 3a: Rooting profile One vegetation-dependent driver of inter-model differences is the
parameterization of root zone soil moisture. Many of these differences are due to the prescribed
distribution of roots in the CLM, which is static in time and varies in space based on PFT. [Hou
et al., 2012] showed that sensitivity depends on the depth of the rooting system, where shallow
root zones tend to be more sensitive to hydrologic variables. In CLM, the water removed from
each soil layer is a function of the effective root fraction within that layer, where the effective
root fraction is weighted by the root contribution from individual PFTs that share the single soil
column. For each PFT, two rooting parameters (ra, rb) are described by [Zeng, 2001]. As part of
our parameter testing, we will examine literature values to determine site-specific rooting depths
for the selected FLUXNET sites and compare these site-based values to those estimated by the
CLM ra/rb scaling. Specifically, we are looking to determine how these rooting profiles change
along the temperate-to-boreal transect in Figure 3, and determine any intra-PFT and climatic
zone dependence to these parameters. As part of this parameter testing, we will design a series
of short-term simulations (restarted from soil moisture equilibrated runs) that test several
different root distributions and evaluate these with SMAP soil moisture and surface fluxes.
Again, we note that different models use very different parameterizations than the one employed
in this single model, but our findings on the role of root zone distribution from this study will
likely cross-over to other models to highlight key spatial (e.g., temperate versus boreal) and
seasonal factors across all model types.
Task 3b: Photosynthesis-stomatal conductance model The use of the Ball-Berry photosynthesis
stomatal conductance model [Collatz et al., 1991] has become the standard method to link ET
and carbon assimilation in most Earth System models. This purely empirical relationship is
described as:
1 𝑚 𝐴 𝑅𝐻
= + 𝑏𝛽!
𝑟! 𝐶!
where rs is the stomatal resistance, m is a dimensionless slope, A is the carbon assimilation rate
(equivalent to GPP when aggregated to the ecosystem scale), RH is the relative humidity at the
leaf surface, Cs is the CO2 concentration at the leaf surface, b is the minimal stomatal
conductance, and βt is the soil water stress function. While several of the parameters (A, RH, Cs)
are calculated online within the model, the slope and intercept (m and b, respectively) are
determined based on values for C3 and C4 plants (m = 9 or 4, and b = 10,000 or 40,000 for C3 and
C4, respectively). The soil water stress, βt, is determined by:

13
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

𝛽! = 𝑤! 𝑟!
!
where for each soil layer i, w is the plant wilting factor and r is the root fraction. In this manner,
βt controls the stomatal conductance and also plays a role in the assimilation rate (A). Its value
ranges from 1 when the soil is wet (e.g., no stress) to 0 when the soil is dry (e.g., high stress),
effectively shutting down transpiration when vegetation is under soil water stress. In this manner,
changes to the root zone structure in Task 2 can also then influence stomatal conductance.
However, here we will focus predominantly on the stomatal resistance empirical parameters of m
and b, and use FLUXNET-based observations of rs (typically calculated at all FLUXNET sites)
to evaluate the modeled calculation of this term. Because of the coupled nature of the
transpiration with soil moisture, we can evaluate the role of the stomatal conductance
parameterization and soil moisture limitations (as constrained by SMAP) on the ET and C fluxes.
Task 3c: Temperature stress Another important driver for understanding water and carbon
fluxes is temperature stress and its effects on vegetation. Sunlit leaves are often a few degrees
Celsius warmer than ambient air temperature, and increased high temperature stress under
climate change may reduce GPP and stomatal closure [Doughty and Goulden, 2008].
The photosynthetic parameters in CLM4.5, including Vcmax and the maximum electron
transport rate (Jmax), leaf nitrogen content, are calculated for an optimum leaf temperature (Tv) of
25°C. To calculate photosynthesis rates at other temperatures, photosynthetic parameters are
scaled by the product of two leaf temperature-dependent function, f(Tv) and fH(Tv):
!"#.!"∆!! ∆!!
!!!"#
∆!! !"#.!" !"#.!"!!"#
𝑓 𝑇! = 𝑒𝑥𝑝 !"#.!"!!"#
1− !!
𝑓! 𝑇! = ∆!!! !∆!!
!!!"#
!!

which increases and decreases with Tv, respectively, such that the product declines about some
optimum temperature. In these equations, Ha is the activation energy, Hd is the deactivation
energy, S is an entropy term, and Rgas is the universal gas constant [8.314 J/molK]. This
optimum is slightly different for each process because the parameters ΔHa, ΔHd, and ΔS vary for
each photosynthetic parameter (Vcmax, Jmax, etc.) [Bonan et al., 2011]. CLM4.5 has the ability for
Vcmax and Jmax to acclimate to temperature changes by modifying the ΔS term, although this
process occurs at timescales longer than the duration of our study [Kattge and Knorr, 2007].
We will investigate whether modifications to ΔHa, ΔHd, and ΔS permit better
parameterization of the temperature dependence of photosynthesis rates (linked to ET rates via
the Ball-Berry formulation). [Bonan et al., 2011] credits the improvement to GPP rates in
CLM4.5 compared to CLM4 in part to the improved temperature dependence of this
parameterization, but we will extend this analysis to test not just gross rates of carbon uptake, but
also against the metrics from Task 3 (e.g., GPP vs soil moisture and ET vs temperature). We
hypothesize that many ESMs simulate high ET rates that prevents carbon uptake in late summer
and that may contribute to a persistent low soil moisture bias.
For Tasks 3a-c, we will identify the most promising configurations based on our parameter
sensitivity tests and identify an optimal configuration for Task 4.

5.4 Task 4: Long term effects of soil moisture-water-carbon-climate feedbacks

Hydrological and carbon feedbacks to climate change will have large impacts on regional
to global scale climate over the next fifty years. Changes in soil moisture, mediated by climate

14
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

change and associated changes in terrestrial ecosystems, will affect regional temperature
extremes [Hirschi et al., 2011]. The impact of water stress is likely to be a primary control on
carbon uptake by terrestrial ecosystems, and thus the amount of anthropogenic CO2 that remains
in the atmosphere to alter Earth’s radiative forcing [Friedlingstein et al., 2013]. This final
synthesis task will extend the temporal scale of Science Question 2 by incorporating new
findings from Tasks 2 and 3 to understand the implications of soil moisture state on long-term
water and carbon fluxes.
We will run three 50-year CESM simulations with CLM4.5-BGC coupled to CAM to
quantify the evolution of soil moisture, carbon uptake, and temperature variability. Each
simulation will use a modified model configuration, including a control run with the standard
CLM4.5-BGC and two simulations where CLM4.5 has been changed according to the parameter
sensitivity tests described in Task 3. We will identify two configurations from Task 3 that best
match both the spatial distribution of SMAP soil moisture and the metrics developed under Task
2 that relate carbon and water fluxes to their environmental drivers. We will use these
simulations to quantify the impact of our parameterizations on the long-term evolution of soil
moisture, as well as the cumulative change to the terrestrial carbon sink owing to more realistic
soil moisture – carbon cycle interactions. Because the seasonal soil moisture and canopy fluxes
govern long-term feedbacks to climate change, we expect that improving the representation at
seasonal timescales will likewise improve our ability to make long-term predictions.
For long-term climate runs, prediction of carbon fluxes enables explicit prediction of the
radiative feedbacks from changing land carbon sinks. We will leverage this capability to assess
how altering the parameterizations that affect soil moisture patterns affect future climate
trajectories. This task will also synthesize our findings on the short time scales (Tasks 1-3) to
understand their implications for the Earth System.

6.0 Project Team, Management and Timeline

6.1 Qualifications of the PIs
The PI (Steiner) and co-Is (Keppel-Aleks and De Roo) bring together a unique set of
capabilities to this project. Steiner is involved in the climate-land surface modeling community
and her publications have explored the key processes in land surface modeling that affect
regional climate [Steiner et al., 2005; Steiner et al., 2009; Tawfik and Steiner, 2011; Bryan et al.,
2015]. Co-I Keppel-Aleks has experience using both remote sensing observations and Earth
system model to understand terrestrial carbon cycling [Keppel-Aleks et al., 2012; Keppel-Aleks
et al., 2013]. She is a member of the Orbiting Carbon Observatory-2 (OCO-2) science team and
of the CESM Biogeochemistry Working Group. She is currently active in the International Land
Model Benchmarking (ILAMB) team, developing open-source metrics for Earth system models
[Mu et al., 2013]. Co-I De Roo brings field-based expertise to the project, and he will assist with
model evaluation using both satellite-derived and ground-based observations. Together, the
project team will synthesize existing knowledge of soil moisture in models with new SMAP
satellite-derived observations to improve our understanding of the role of soil moisture on
surface water and carbon fluxes.

6.2 Project management

The full project team will be located at the University of Michigan (UM) with overall
project management by PI Steiner. At UM, she will work with the two co-Is on the project
(Keppel-Aleks, De Roo) and the postdoctoral fellow to meet project milestones. Full team
meetings will occur quarterly to discuss the project progress. Because all team members are

15
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

located within the same department at UM, this will enable convenient meeting times with
minimal travel costs. One postdoctoral fellow will be hired to work on the project, and will work
closely with Steiner, Keppel-Aleks, and De Roo (see Table 2). Results will be disseminated
through participation in annual SMAP Science Team meetings, attendance at national
conferences such as the American Geophysical Union (AGU), and peer-reviewed publications.

6.3 Timeline
Table 2 shows the proposed timeline of the project as well as individual team member’s
responsibilities. With a start date of 1 April 2016, we anticipate that the first full year of SMAP
data (including one full growing season plus portions of the second year) will be available in
year 1 of the project. Task 1 will take place in year 1, including understanding the data
constraints of the vegetation water content (VWC) criteria and developing novel techniques to
use SMAP data in off-seasons for deciduous forested regions (Task 1a), as well as evaluation of
the modeled soil moisture state (Task 1b) and evaluation of modeled soil moisture and SMAP
versus ground-based observations (Task 1c). At the end of Year 1, we will begin Task 2,
including developing an understanding of observed and modeled soil moisture-water-carbon
feedbacks. In years 2 and 3, we will conduct the parameter testing of Task 3. In the final year of
the project (year 3), we will extend our results to longer-term simulations and identify key
features that may affect the regional hydroclimate as well as long-term carbon storage (Task 4).
We have built in overlap in the timeline to account for changes in the work plan schedule, to
allow for lags in data availability. We have also developed a contingency plan with regard to the
VWC criteria of Task 1a, and plan to focus on deciduous broadleaf forests in the event that we
are unable to determine off-season data for boreal regions. Due to the experience of the project
team, we will be able to quickly adapt to these data limitations and still meet the objectives of the
proposal.

Table 2. Project timeline and responsibilities. (PD = postdoctoral fellow)

Year 1 Year 2 Year 3
2016 2017 2018 2019
Project Task Participants S S F W S S F W S S F W
Task 1: Evaluation of modeled SM state
- Task 1a: VWC constraints Steiner, Keppel-Aleks, deRoo, PD
- Task 1b: Evaluation of modeled SM state Steiner, PD
- Task 1c: Evaluation with ground-based obs Steiner, deRoo, PD
Task 2: Process-level energy-H2O-CO2
- Task 2a: ET metrics Steiner, PD
- Task 2b: Carbon metrics Keppel-Aleks, PD
- Task 2c: Precipitation recycling Steiner, PD
Task 3: Parameter testing
- Task 3a: Root zone distributions Steiner, deRoo, PD
- Task 3b: Photosynthesis - rs models Steiner, Keppel-Aleks, PD
- Task 3c: Temperature stress Keppel-Aleks, PD
Task 4: Long-term effects
- Model simulations PD
- Analysis Steiner, Keppel-Aleks, deRoo, PD

16
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

REFERENCES
Baldocchi, D. D., et al. (2001), FLUXNET: A New Tool to Study the Temporal and Spatial
Variability of Ecosystem-Scale Carbon Dioxide, Water Vapor and Energy Flux Densities,
Bulletin of the American Meteorological Society 82, 2415-2435.
Bonan, G. B., K. W. Oleson, R. A. Fisher, G. Lasslop, and M. Reichstein (2012), Reconciling
leaf physiological traits and canopy flux data: Use of the TRY and FLUXNET databases
in the Community Land Model version 4, J. Geophys. Res.-Biogeosci., 117.
Bonan, G. B., P. J. Lawrence, K. W. Oleson, S. Levis, M. Jung, M. Reichstein, D. M. Lawrence,
and S. C. Swenson (2011), Improving canopy processes in the Community Land Model
version 4 (CLM4) using global flux fields empirically inferred from FLUXNET data, J.
Geophys. Res.-Biogeosci., 116.
Bryan, A. M., A. L. Steiner, and D. J. Posselt (2015), Regional modeling of surface-atmosphere
interactions and their impact on Great Lakes hydroclimate, Journal of Geophysical
Research-Atmospheres, 120(3), 1044-1064.
Chan, S., R. Bindlish, R. Hung, T. Jackson, and J. Kimball (2013), Vegetation Water Content,
Ancillary ReportRep., Jet Propulsion Laboratory, Pasadena, CA.
Clapp, R. B., and G. M. Hornberger (1978), EMPIRICAL EQUATIONS FOR SOME SOIL
HYDRAULIC-PROPERTIES, Water Resources Research, 14(4), 601-604.
Collatz, G. J., M. Ribas-Carbo, and J. A. Berry (1992), COUPLED PHOTOSYNTHESIS-
STOMATAL CONDUCTANCE MODEL FOR LEAVES OF C4 PLANTS, Australian
Journal of Plant Physiology, 19(5), 519-538.
Collatz, G. J., J. T. Ball, C. Grivet, and J. A. Berry (1991), Physiological and environmental
regulation of stomatal conductance, photosynthesis, and transpiration: A model that
includes a laminar boundary layer. , Agricultural and Forest Meteorology, 54, 107-136.
Davidson, E. A., and I. A. Janssens (2006), Temperature sensitivity of soil carbon decomposition
and feedbacks to climate change, Nature, 440(7081), 165-173.
Dirmeyer, P. A., C. A. Schlosser, and K. L. Brubaker (2009), Precipitation, Recycling, and Land
Memory: An Integrated Analysis, Journal of Hydrometeorology, 10(1), 278-288.
Doughty, C. E., and M. L. Goulden (2008), Are tropical forests near a high temperature
threshold?, Journal of Geophysical Research-Atmospheres, 113, G00B07.
Farquhar, G. D., S. von Caemmerer, and J. A. Berry (1980), A biochemical model of
photosynthetic CO2 assimilation in leaves of C3 species, Planta, 149, 78-90.
Frankenberg, C., C. O'Dell, J. A. Berry, L. Guanter, J. Jointer, P. Kohler, R. Pollack, and T. E.
Taylor (2014), Prospects for chlorophyll fluoresence remote sensing from the Orbiting
Carbon Observatory-2, Remote Sens. Environ., 147, 1-12.
Friedlingstein, P., M. Meinshausen, V. K. Arora, C. D. Jones, A. Anav, S. K. Liddicoat, and R.
Knutti (2013), Uncertainties in CMIP5 Climate Projections due to Carbon Cycle
Feedbacks, Journal of Climate, 27, 511-526.
Gent, P. R., et al. (2011), The Community Climate System Model Version 4, Journal of Climate,
24(19), 4973-4991.
Hirschi, M., S. I. Seneviratne, V. Alexandrov, F. Boberg, C. Boroneant, O. B. Christensen, H.
Formayer, B. Orlowsky, and P. Stephanek (2011), Observational evidence for soil-
moisture impact on hot extremes in southeastern Europe, Nature Geoscience, 4, 17-21.
Hou, Z., M. Huang, R. Leung, G. Lin, and D. M. Ricciuto (2012), Sensitivity of surface flux
simulations to hydrologic parameters based on an

17
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

uncertainty quantification framework applied to the Community Land Model,, Journal of

Geophysical Research-Atmospheres, 117(D15108).
Hurrell, J. W., et al. (2013), The Community Earth System Model A Framework for
Collaborative Research, Bulletin of the American Meteorological Society, 94(9), 1339-
1360.
Jung, M., et al. (2010), Recent decline in the global land evapotranspiration trend due to limited
moisture supply, Nature, 467(7318), 951-954.
Kattge, J., and W. Knorr (2007), Temperature acclimation in a biochemical model of
photosynthesis: a reanalysis of data from 36 species, Plant Cell Environ., 30, 1176-1190.
Keppel-Aleks, G., A. S. Wolf, M. Mu, S. C. Doney, D. C. Morton, P. S. Kasibhatla, J. B. Miller,
E. J. Dlugokencky, and J. T. Randerson (2014), Separating the influence of temperature,
drought, and fire on interannual variability in atmospheric CO2, Global Biogeochemical
Cycles, 28(11), 1295-1310.
Keppel-Aleks, G., et al. (2013), Atmospheric Carbon Dioxide Variability in the Community
Earth System Model: Evaluation and Transient Dynamics during the Twentieth and
Twenty-First Centuries, Journal of Climate, 26(13), 4447-4475.
Keppel-Aleks, G., et al. (2012), The imprint of surface fluxes and transport on variations in total
column carbon dioxide, Biogeosciences, 9, 875-891.
Koster, R. D., et al. (2004), Regions of strong coupling between soil moisture and precipitation,
Science, 305(5687), 1138-1140.
Lawrence, D. M., and A. G. Slater (2008), Incorporating organic soil into a global climate
model, Climate Dynamics, 30(2-3), 145-160.
Lee, J.-E., et al. (2013), Forest productivity and water stress in Amazonia: Observations from
GOSAT chlorophyll fluorescence, Proceedings of the Royal Society B: Biological
Sciences(280), 1761.
Lei, H., M. Huang, L. R. Leung, D. Yang, X. Shi, J. Mao, D. J. Hayes, C. R. Schwalm, Y. Wei,
and S. Liu (2014), Sensitivity of global terrestrial gross primary production to hydrologic
states simulated by the Community Land Model using two runoff parameterizations,
Journal of Advances in Modeling Earth Systems, 6(3), 658-679.
Lindsay, K., G. B. Bonan, S. C. Doney, F. M. Hoffman, D. M. Lawrence, M. C. Long, N. M.
Mahowald, J. K. Moore, J. T. Randerson, and P. E. Thornton (2014), Preindustrial-
Control and Twentieth-Century Carbon Cycle Experiments with the Earth System Model
CESM1(BGC), Journal of Climate, 27(24), 8981-9005.
Lofgren, B. M., A. D. Gronewold, A. Acciaioli, J. Cherry, A. Steiner, and D. Watkins (2013),
Methodological Approaches to Projecting the Hydrologic Impacts of Climate Change,
Earth Interactions, 17.
Matheny, A. M., et al. (2014), Characterizing the diurnal patterns of errors in the prediction of
evapotranspiration by several land-surface models: An NACP analysis, J. Geophys. Res.-
Biogeosci., 119(7), 1458-1473.
Mitchell, K. E., et al. (2004), The multi-institution North American Land Data Assimilation
System (NLDAS): Utilizing multiple GCIP products and partners in a continental
distributed hydrological modeling system, Journal of Geophysical Research-
Atmospheres, 109,(D07S90).
Mu, M., F. M. Hoffman, D. M. Lawrence, W. J. Riley, G. Keppel-Aleks, and J. T. Randerson
(2013), Design and development of a community carbon cycle benchmarking system for

18
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

CMIP5 models, paper presented at American Geophysical Union Fall Meeting, San
Francisco, CA.
Mu, Q., M. Zhao, and S. W. Running (2011), Improvements to a MODIS global terrestrial
evapotranspiration algorithm, Remote Sens. Environ., 115(8), 1781-1800.
Neale, R. B., J. Richter, S. Park, P. H. Lauritzen, S. J. Vavrus, P. J. Rasch, and M. Zhang (2013),
The Mean Climate of the Community Atmosphere Model (CAM4) in Forced SST and
Fully Coupled Experiments, Journal of Climate, 26(14), 5150-5168.
Niu, G. Y., and Z. L. Yang (2006a), Effects of frozen soil on snowmelt runoff and soil water
storage at a continental scale, Journal of Hydrometeorology, 7, 937-952.
Niu, G. Y., and Z. L. Yang (2006b), Effects of frozen soil on snowmelt runoff and soil water
storage at a continental scale, Journal of Hydrometeorology, 7, 937-952.
Oleson, K. W., D. M. Lawrence, and e. al. (2013), Technical Description of version 4.5 of the
Community Land Model (CLM)Rep., 434 pp, National Center for Atmospheric Research,
Boulder, CO.
Parton, W. J., D. S. Schimel, C. V. Cole, and D. S. Ojima (1987), Analysis of factors controlling
soil organic matter levels in Great Plains Grasslands, Soil Sci. Soc. Am. J., 51, 1173-1179.
Saleska, S. R., et al. (2003), Carbon in amazon forests: Unexpected seasonal fluxes and
disturbance-induced losses, Science, 302(5650), 1554-1557.
Schar, C., D. Luthi, U. Beyerle, and E. Heise (1999), The soil-precipitation feedback: A process
study wtih a regional climate model, Journal of Climate, 12(3), 722-741.
Sellers, P. J., D. A. Randall, G. J. Collatz, J. A. Berry, C. B. Field, D. A. Dazlich, C. Zhang, G.
D. Collelo, and L. Bounoua (1996), A revised land surface parameterization (SiB2) for
atmospheric GCMs .1. Model formulation, Journal of Climate, 9(4), 676-705.
Seneviratne, S. I., T. Corti, E. L. Davin, M. Hirschi, E. B. Jaeger, I. Lehner, B. Orlowsky, and A.
J. Teuling (2010), Investigating soil moisture-climate interactions in a changing climate:
A review, Earth-Science Reviews, 99(3-4), 125-161.
Steiner, A., J. Pal, F. Giorgi, R. Dickinson, and W. Chameides (2005), The coupling of the
Common Land Model (CLM0) to a regional climate model (RegCM), Theoretical and
Applied Climatology, 82(3-4), 225-243.
Steiner, A. L., J. S. Pal, S. A. Rauscher, J. L. Bell, N. S. Diffenbaugh, A. Boone, L. C. Sloan, and
F. Giorgi (2009), Land surface coupling in regional climate simulations of the West
African monsoon, Climate Dynamics, 33(6), 869-892.
Suyker, A. E., S. B. Verma, and G. G. Burba (2003), Interannual variability in net CO2 exchange
of a native tallgrass prairie, Glob. Change Biol., 9(2), 255-265.
Tawfik, A. B., and A. L. Steiner (2011), The role of soil ice in land-atmosphere coupling over
the United States: A soil moisture-precipitation winter feedback mechanism, Journal of
Geophysical Research-Atmospheres, 116.
Taylor, K. E., R. J. Stouffer, and G. A. Meehl (2012), AN OVERVIEW OF CMIP5 AND THE
EXPERIMENT DESIGN, Bulletin of the American Meteorological Society, 93(4), 485-
498.
Thornton, P. E., J.-F. Lamarque, N. A. Rosenbloom, and N. M. Mahowald (2007), Influence of
carbon-nitrogen cycle coupling on land model response to CO2 fertilization and climate
variability, Global Biogeochemical Cycles, 21(4).
Wang, W., P. Ciais, R. R. Nemani, J. G. Canadell, S. Piao, S. Sitch, M. A. White, H. Hashimoto,
C. Milesi, and R. B. Myneni (2013), Variations in atmospheric CO2 growth rates coupled

19
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

with tropical temperature, Proceedings of the National Academy of Sciences of the United
States of America, 110(32), 13061-13066.
Weiss, M., and L. Menzel (2008), A global comparison of four potential evapotranspiration
equations and their relevance to stream flow modeling in semi-arid environments,
Advances in Geosciences, 18, 15-23.
Xia, Y., M. T. Hobbins, Q. Mu, and M. B. Ek (2015), Evaluation of NLDAS-2
evapotranspiration against tower flux site observations, Hydrological Processes, 29(7),
1757-1771.
Xia, Y., et al. (2012), Continental-scale water and energy flux analysis and validation for the
North American Land Data Assimilation System project phase 2 (NLDAS-2): 1.
Intercomparison and application of model products, Journal of Geophysical Research-
Atmospheres, 117.
Zeng, X., and M. Decker (2009), Improving the Numerical Solution of Soil Moisture-Based
Richards Equation for Land Models with a Deep or Shallow Water Table, Journal of
Hydrometeorology, 10(1), 308-319.
Zeng, X. B. (2001), Global vegetation root distribution for land modeling, Journal of
Hydrometeorology, 2(5), 525-530.

20
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

CVs

21
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

SUMMARY OF WORK EFFORT

Summary of Work Effort

NAME TITLE Yr 1 Yr 2 Yr 3 Total

Allison Steiner Principal Investigator 0.7 0.7 0.7 2.1
Gretchen Keppel-Aleks Co-Investigator 0.7 0.7 0.7 2.1
Roger DeRoo Co-Investigator 0.2 0.2 0.2 0.7
TBD Postdoctoral Research Fellow 12.0 12.0 12.0 36
13.6 13.6 13.6 40.9

22
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

CURRENT AND PENDING SUPPORT

23
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

BUDGET JUSTIFICATION

24
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

FACILITIES AND EQUIPMENT

The scientific research will be conducted on the campus of the University of Michigan (UM) in
Ann Arbor, MI. All investigators on the project (Steiner, Keppel-Aleks, De Roo) and the
postdoctoral fellow will be located in the Climate and Space Sciences and Engineering (CLaSP)
building on the North Campus location of the University of Michigan.

Office: Existing office and library resources will be utilized for the investigators and the
postdoctoral fellows at no cost to the project.

Computer modeling: High performance computing will be conducted using the Flux cluster at the
University of Michigan Advanced Research Computing. Flux is an HPC Linux-based cluster
intended to support parallel and other applications that are not suitable for departmental or
individual computers. Each Flux compute node comprises multiple CPU cores with at least 4 GB
of RAM per core; Flux has more than 19,000 cores. All compute nodes are interconnected with
InfiniBand networking. Computing jobs on Flux are managed through a combination of the
Moab Scheduler, the Terascale Open-Source Resource and QUEue Manager (Torque) and the
GOLD Allocation Manager from Adaptive Computing. The system also includes high-speed
scratch storage using the Lustre parallel network file system. The storage is connected with
InfiniBand. This file system allows researchers to store data on a short-term basis to perform
calculations; it is not for long-term data storage or archival purposes. All Flux nodes are
interconnected with quad-data rate InfiniBand, delivering up to 40 Gbps of bandwidth and less
than 5µs latency. Flux is connected to the University of Michigan’s campus backbone to provide
access to student and researcher desktops as well as other campus computing and storage
systems. The campus backbone provides 100 Gbps connectivity to the commodity Internet and
the research networks Internet2 and MiLR. Flux is housed in the Modular Data Center (MDC).
The MDC uses ambient air for cooling approximately 75% of the year, thus significantly
reducing the amount of energy needed for cooling, and contributing to U-M’s sustainability
efforts. Flux computing services are provided through a collaboration of University of Michigan
units: Advanced Research Computing (in the Office of the VP of Research and the Provost’s
Office), and computing groups in schools and colleges at the university. For this project,
dedicated nodes and storage space will be purchased on the cluster (see “Other Direct costs,”
above) to ensure constant availability for project simulations.

The PI Steiner has an additional Linux server available for the research group at no cost for data
transfer, post-processing and analysis. Students will use existing laptops for individual
workstations, and use the Flux computing cluster and value-added data storage for simulations
and storage of model and SMAP data.

25
 NASA SMAP Proposal
Steiner, Keppel-Aleks & De Roo

Budget Detail

26