Developing a strategy for the national coordinated soil moisture monitoring network

Soil moisture is a critical land surface variable, affecting a wide variety of climatological, agricultural, and hydrological processes. Determining the current soil moisture status is possible via a variety of methods, including in situ monitoring, remote sensing, and numerical modeling. Although all of these approaches are rapidly evolving, there is no cohesive strategy or framework to integrate these diverse information sources to develop and disseminate coordinated national soil moisture products that will improve our ability to understand climate variability. The National Coordinated Soil Moisture Monitoring Network initiative has developed a national strategy for network coordination with NOAA's National Integrated Drought Information System. The strategy is currently in review within NOAA, and work is underway to implement the initial milestones of the strategy. This update reviews the goals and steps being taken to establish this national‐scale coordination for soil moisture monitoring in the United States.


INTRODUCTION
Soil moisture is a critical land surface variable affecting a wide variety of economically and environmentally important processes. From agricultural monitoring, to weather prediction, to drought and flood mitigation, the value of soil moisture metrics is undeniable (Vereecken et al., 2008). Most ground-based networks use in situ sensors measuring at high temporal resolution and multiple soil depths, but the volume of measurement is typically small. Remote sensing platforms have much larger spatial footprints (10-40 km) but only sense shallow soil moisture (<5 cm) with return periods every 2-3 d.
Lastly, land surface models (LSMs) can estimate soil moisture with high spatial and temporal resolution, but they are imperfect approximations of the real-world physics that rely on meteorological data and underlying parameterizations. In fact, both space-borne and LSM estimates of soil moisture require calibration and validation to in situ, ground validation data. Thus, these three sources of data are required to work in concert to produce a temporally and spatially continuous soil moisture product at the relevant scale needed.
The United States has a prolific, but uncoordinated, collection of in situ monitoring networks at the national, state, and local levels ( Figure 1). However, there is currently no national strategy for the development, deployment, and maintenance of these soil moisture monitoring networks, nor for their coordination and data integration. The absence of such a strategy leads to a host of problems including inadequate monitoring in many states, inconsistent data collection practices between networks, and no cohesive plan to improve the overall infrastructure. Here, we summarize a coherent strategy for the National Coordinated Soil Moisture Monitoring Network (NCSMMN), developed for the National Integrated Drought Information System (NIDIS) under the NOAA, the entity tasked by Congress to manage this initiative. This update presents the key components of this strategy, results from the associated 2020 National Soil Moisture Workshop, and a path forward for the NCSMMN.

In situ soil moisture sensors
Soil moisture is usually measured as volumetric soil water content (SWC) or the volume of liquid water within a given volume of soil (m 3 m −3 ). Soil water content can range from oven dry (0 m 3 m −3 ) to the water-filled porosity of a saturated soil, typically <0.60 m 3 m −3 . Most soil moisture sensors infer SWC from either thermal or electrical properties of the bulk soil; the latter tends to be more popular due to the wider availability of commercial sensors and perceived simplicity of the measurement. Most electrical SWC sensors are based on the propagation of an electromagnetic wave within a porous medium. These fall into many different classes including time domain reflectometry, time domain transmissometry, transmission line oscillators, capacitance sensors, and impedance sensors (Vaz et al., 2013). Measurement errors estimated by manufacturers under carefully controlled conditions are often 0.02-0.03 m 3 m −3 , but errors estimated by researchers in field and laboratory experiments are often substantially higher (Table 1). However, these measurement errors can be reduced through improved, and often site-specific field or laboratory calibrations. Ultimately, the soil moisture measurements from in situ networks should be validated using volumetric soil sampling at each station to determine the ground validation values and network-level measurement error, but few in situ networks have been validated to date (Caldwell et al., 2019;Coopersmith et al., 2015;Scott et al., 2013;Zhang et al., 2019). Currently, there are no standard or widely accepted methods for installation, calibration, validation, and quality control for SWC sensors. This lack of standardization and general guidance has made it challenging for some monitoring networks, like state Mesonets, to add soil moisture measurements.

Remote sensing platforms
Space-borne microwave soil moisture sensors can either be passive (receive energy) or active (transmit and receive energy). Passive remote sensors (radiometers) measure brightness temperature emissions from microwave radiation originating from the Earth's surface. The frequency and intensity of emitted radiation depends on the dielectric properties of the near surface, which for soils are a function of the amount of water present and its temperature. Active remote sensors (or radars) provide their own illumination source, sending out a transmitted wave and measuring the received reflection back to determine its backscatter cross-section. Synthetic aperture radars use processing that provides higher

Core Ideas
• Soil moisture is a critical yet underrepresented land surface variable. • Soil moisture data collection is undergoing rapid growth and innovation. • We present a strategy for a nationally coordinated monitoring network.
spatial resolution, allowing finer scale features of the surface to be observed. Measurements of emissivity and backscatter cross-section (or simply backscatter) provide complementary information on the soil moisture, surface roughness, and vegetation characteristics of the land surface (see Tables 2 and  3). Reviews of various satellite-based soil moisture platforms and related issues can be found in Mohanty et al. (2017) and Babaeian et al. (2019). An ultimate goal of NCSMMN would be to have quality standards that are comparable with the Fiducial Reference Measurement (FRM) standard, as implemented at https://qa4sm.eu/.

Land surface models
Land surface models are systems of equations designed to simulate the flow of mass, water, and energy within the soil-vegetation-atmosphere continuum. The water balance approach applied by LSM calculates a change in soil water storage as the difference between incoming (e.g., precipitation) and outgoing (e.g., evapotranspiration, runoff, and groundwater recharge) fluxes of water. Land surface models differ widely with regards to their complexity, assumptions, and atmospheric forcing requirements. Model-based soil moisture datasets are easily accessible and provide temporal continuity (i.e., no missing data) and continuous spatial coverage within their simulation domain. However, LSMs have several key limitations for soil moisture including simplified physics  and inadequate parameterization schemes for soil properties (Fatichi et al., 2020). In addition, LSM performance and accuracy are highly susceptible to the quality of the forcing data, including precipitation, temperature, net radiation, humidity, and wind. The large availability of routinely delivered forcing data, along with the long-term trend in computational power, has substantially reduced obstacles for operational, large-scale soil moisture products derived from LSM (Tables 2 and 3). For a review of regional and global land data assimilation systems, see Xia et al. (2019).
T A B L E 1 A summary of common (but not all-inclusive) in situ and profile sensor errors, as RMSE, stated from the manufacturer and determined by researchers using the factory standard coefficients and soil-specific calibrations.

CURRENT STATE OF SOIL MOISTURE MONITORING IN THE USA
The number of in situ soil moisture monitoring stations has increased substantially in recent decades. In the United States, most long-term soil moisture monitoring networks are operated by federal and state agencies. These networks have continued to expand and infill at both regional and national scales. Figure 1 provides an overview of key federal, state, and university-sponsored networks currently in operation with data transmitted in near real time. Some of these networks have a period of record beyond 20 yr; however, there is also a substantial variability in soil depths monitored and type of sensors used ( Table 4). As of 2021, there are ∼2,000 soil moisture monitoring stations producing publicly available data in the United States.

DEVELOPING A STRATEGY FOR THE NCSMMN
In 2013, NIDIS and partners began an initiative to work towards a coordinated national soil moisture monitoring network. A meeting to clarify the vision for this effort was held in November 2013 in Kansas City, MO, with federal, state, and academic experts participating (McNutt et al., 2013). A , an Executive Committee that included leaders from federal agencies and academic institutions was formed and was charged with clearly defining the goals and framework to bring the NCSMMN concept to fruition (Clayton et al., 2019). Drawing on knowledge and data generated from this series of meetings and associated research projects, the Executive Committee, working with other partners, prepared a "A Strategy for the National Coordinated Soil Moisture Monitoring Network," which is summarized below.

OVERVIEW OF THE NCSMMN STRATEGY
The NCSMMN is a multi-institutional national effort with the mission to provide "coordinated, high-quality, nationwide, soil moisture information for the public good." At the highest level, the NCSMMN seeks to • establish a national "network of networks" that effectively demonstrates data and operational coordination of in situ networks, such as those shown in Table 4, and addresses gaps in coverage; • support research and development of innovative techniques to merge in situ soil moisture data with remotely sensed and modeled hydrologic data to create near-real-time, gridded, user-friendly soil moisture maps and associated tools; and • build a community of practice and expertise around measuring soil moisture and developing new ways to use soil moisture information-a "network of people" that links data providers, researchers, and the user community.
The Strategy Document for the NCSMMN presents several recommendations and next steps for moving these goals forward. The recommendations are summarized in Table 5 and listed in a logical flow of activities, but many steps are intended to be taken in parallel. The first group of recommendations address NCSMMN operations and support activities, including determining a formal institutional "home" for the NCSMMN and engaging in communication and outreach. Currently, NIDIS is serving as the lead agency for the NCSMMN and has developed an initial NCSMMN webpage on its drought portal (https://www.drought.gov/droughtin-action/national-coordinated-soil-moisture-monitoringnetwork). An NCSMMN email listserv has also been established, and we invite interested individuals to sign up using information provided on the webpage. Other outreach Develop, release, and promote NCSMMN products activities include a series of workshops and seminars planned for the coming year, including a Mesonet operators' workshop to provide peer-to-peer networking (see the NCSMMN webpage for more details on outreach activities). A second area of focus in the NCSMMN Strategy is on developing the appropriate infrastructure for high-quality data integration. Accordingly, recommendations in the Strategy aim to formalize and codify partnerships with existing state Mesonets, as well as to develop quality criteria for data inclusion. Another recommendation is to increase the density of networks nationwide through targeted build-outs, and by exploring potential new partnerships, including private sector and citizen science efforts.
The final area of focus in the NCSMMN Strategy is on product development. To deliver the intended products to support public decision-making, the Strategy recommends supporting research to develop or improve methodologies for soil moisture data collection, standardization, integration, blending, and validation. One example is the issue of how best to perform interpolation (horizontal, vertical, temporal) of point source data into meaningful gridded information. The final recommendation is to develop products that meet the needs of diverse end-user groups, and that support crucial applications such as drought and flood monitoring, fire danger ratings, and streamflow forecasting.

COMMUNITY INPUT ON THE NCSMMN STRATEGY
The 2020 National Soil Moisture Workshop was held online on 12-13 August, with 182 attendees from federal, state, and local agencies; universities; and the private sector. This annual workshop provides a unique opportunity for leaders in soil moisture research and development to come together in an interactive format to exchange ideas and develop collaborations. This was the 10th consecutive year for this workshop. One objective of this year's workshop was to gather additional community input on the NCSMMN strategy and to stimulate progress towards realizing the vision of the NCSMMN.
Participants were assigned breakout groups to give feedback on the NCSMMN Strategy through a series of three overarching topics summarized in Figure 2 and elaborated upon here. Because a "network of networks" requires some assessment of data quality from each provider to properly assign weight to that data in generated products, our first topic focused on establishing data quality criteria. We asked: What criteria should be used to assess "high-quality" (or Tier 1) versus "moderate-quality" (Tier 2) data? Metadata, the data behind the data, was considered to be of particular importance and in fact has been a recurring theme in NCSMMN discussions. Different types of metadata are listed in Figure 3. One key type of metadata is soil characterization for each loca-tion and measurement depth. Tier 1 data providers should also provide raw data values along with sensor calibrations and some measure of network error and uncertainty, and have documented quality assurance/quality control ideally with redundancy in measurements. A basic requirement for a NCSMMN provider is access to data with minimal latency, which necessitates automated quality assurance flagging to assess abrupt changes or steps. Most modern soil moisture sensors also collect temperature and bulk electrical conductivity data. These data, along with ancillary time series data from meteorological sensors, and site cameras, would also improve the overall quality and confidence in the data provided. It should be noted that network quality may not be constant in either space or time due to factors such as discontinuity in funding and locations subjected to deposition, erosion, biota, and expansive soils, all of which can change readings.
Our second breakout topic was an exploration of impediments to and user needs for data quality: What are the technical or other (e.g., organizational) impediments to generating high-quality data? And what technical assistance is needed to help data providers deliver high-quality data? The foremost response was financial support. In most organizations, it is easier to acquire initial funds to purchase equipment or install a network than long-term funds for operations and maintenance. Second was technical support. Given a general absence of standards, limited number of qualified staff, and lack of institutional expertise, training programs and working groups are needed to assist network operators with installation, maintenance, data transmission, and quality control. Data management and dissemination at some final repository is needed, perhaps along the lines of the National Ground Water Monitoring Network (NGWMN), which is a compilation of selected groundwater monitoring wells from federal, state, and local groundwater monitoring network (SOGW, 2013). Data ownership and network identity were also noted as impediments, since many data producers are required to show usage and benefits to justify their costs.
In regard to NCSMMN data outputs, we asked: What are the most important data attributes or products to meet user needs? The community responses highlighted data availability, focusing on gap-filled time series data for a uniform set of measurement depths in a consistent format, along with interactive charts and web applications. For spatially interpolated (i.e., gridded) data, color-indexed maps with daily, weekly, or monthly summaries (not raw data) were requested. The requested data formats included time synched, station time series data, and GeoTIFF or netCDF for gridded products, which tend to be cloud friendly, as files become large. Some decision making requires near real-time data for emergency management, flood forecasting, agricultural applications (irrigation requirements, fertilizer and pesticide applications, harvesting and planting decisions), and wildfire potential and fuel moisture estimation. The requested  Lastly, many users do not have technical expertise interpreting soil conditions, so some level of education and outreach is required. Technical workshops on topics such as data use, products, and the latest technologies in sensors would improve usage of any of these products.
The final breakout sessions focused on NCSMMN research priorities and products in the near and long term. The immediate needs included an effective Data Management Plan; developing a repository of data processing scripts; conducting a national assessment of networks to determine where spatial coverage is either lacking or redundant; advancing approaches to soil moisture measurement in more complex terrains such as forests, alpine terrain of varied aspect/slope, and under rainfed and irrigated crops; and convening a steering committee. Many of these near-term priorities are currently being addressed as noted in Section 7. In the long term, it was stated that the NCSMMN should prioritize network expansion to increase the overall density of data. The last major priority is to expand the soil moisture community to include other sciences such as social sciences and economics, human health, soil health, and so on, and continue to improve collaboration between data providers, researchers, and nonresearch data users through webinars and workshops.

National soil survey participation
It has been recognized that information about the soil (<2 m) and vadose zone (the entire unsaturated zone) is critical to the interpretation of any remote sensing or LSM product. To support this crucial collateral information, the Kellogg Soil Survey Laboratory in Lincoln, NE, is eager to support the analysis and archiving of soil samples collected at monitoring station locations to improve their soil archive, as well as to provide the necessary metadata for each station. A minimum set of soil parameters are to be determined for each soil moisture station by providing soil cores to the Kellogg Laboratory for analysis.

Installation guidance
As noted above, there is a need for formal guidance on site selection and soil moisture sensor installation. Building off of the IAEA (2008) Training Course Series, the USGS plans to produce a collaborative Techniques and Methods (T&M) guide on soil moisture data collection. The USGS T&M series compiles the description of procedures for the collection, analysis, or interpretation of scientific data. It includes selected scripts, manuals, and documentation that represent major methodology or techniques of data collection. In conjunction with USDA-ARS, the USGS is updating a former T&M on soil moisture by Johnson (1962) to serve as a handson installation guide for field technicians. Drawing off this work, the NCSMMN Executive Committee is planning to develop a video guide for sensor installation along the lines of the Lawrence et al. (2016) approach to sampling forest soils.

NCSMMN web presence
As mentioned, NIDIS has developed an initial web presence for NCSMMN communication and public outreach, with plans to broaden this platform over time as the NCSMMN organizational alignment becomes more settled. In addition, an Open Science Framework project has been established (https://osf.io/56gsj/) to serve as a resource for the Executive Committee and for community interaction. This site provides a repository for committee deliberations and includes various background documents related to the NCSMMN.

Upcoming workshops
One of the primary purposes of the NCSMMN is to provide engagement across the many different groups using or generating soil moisture data. As such, a critical method of engagement is workshops and seminars to promote conversations and sharing of knowledge. A sequence of workshops and seminars are now in the planning stages. The Soil Moisture Network Operators Workshop (SM-NOW) will serve as a data provider support forum for peer-to-peer sharing of techniques and experiences to help improve the installation, maintenance, and data delivery from soil moisture networks. This community is expected to benefit from internal discussions of siting strategies, management protocols, and other challenges faced by network operators and managers. A series of Soil Moisture End Users Workshops are being planned to provide an opportunity for different soil moisture data end user sectors (such as state climatologists, water basin managers, drought monitor authors, weather forecasters, etc.) to provide specific ideas and needs they have for useful soil moisture products. The objective is to create a more tailored and detailed set of user needs, to better inform and orient research and product development efforts. For example, a workshop focused on the relationships between soil moisture and wildfire danger is being planned for spring 2021. A seminar series is also being organized to provide more regular, less time-demanding updates for the soil moisture community on new research and project developments. This is currently planned for quarterly calls (four per year) with one being synchronous with the National Soil Moisture Workshop. For more information on any of these workshops or seminars, contact the corresponding author.

OTHER RELATED ACTIVITIES
The validation of global coarse satellite soil moisture products requires a community-based effort to implement best practices (Gruber et al., 2020). The Committee on Earth Observation Satellites (CEOS) has the goal of ensuring international coordination of civil space-based Earth observation programs, promoting exchange of data to optimize societal benefit and to inform decision making for securing a prosperous and sustainable future for humankind. The mission of the Working Group on Calibration and Validation is to ensure the accuracy and quality of Earth Observation data and products. The CEOS Land Product Validation Soil Moisture Subgroup recently authored the Soil Moisture Product Validation Good Practices Protocol (Montzka et al., 2020) to provide, analyze, and improve high quality Earth Observation results; to evaluate the long-term quality of soil moisture products; to give advice on how to handle temporal and spatial mismatch; and to provide guidance on effectively reporting validation results.
As mentioned previously, the U.S. Army Corps of Engineers has begun the process of awarding contracts to state and federal agencies, as well as private firms, to expand the monitoring of soil moisture and snowpack in the Upper Missouri River basin (USACE, 2021). These contracts are expected to increase the number of public monitoring stations in the basin by approximately 540 sites and will take 5-7 yr to complete. It is anticipated that this expansion will provide better input data for basin runoff models and better inform decision making for hydrologic concerns in the basin as well as downstream. More generally, data from the expansion will be integrated into the overall NCSMMN initiative and support a broad range of research efforts and decision-making applications related to flooding, drought, water and weather forecasting.
Recently, the USGS has begun integrating its water science programs to better address the nation's greatest water resource challenges now and into the future by advancing data collection in 10 prioritized basins (Van Metre et al., 2020). Three new programs instrumental in launching this basin selection effort are the Next Generation Water Observing Systems (NGWOS), Integrated Water Availability Assessments (IWAA), and Integrated Water Prediction (IWP). Under NGWOS, traditional USGS hydrologic data, including river discharge and groundwater levels, will be increasingly collected using more advanced and novel collection methods to improve modeling and prediction capabilities. Additionally, other aspects of the hydrologic cycle, primarily evapotranspiration, snowpack, and soil moisture, will be included to support both IWAA and IWP programs, as well as to provide real-time data to national and regional modeling efforts and the NCSMMN. Instrumentation testing and deployment began in 2018 in the Delaware River basin as part of a pilot effort and will be enhanced in the Upper Colorado and Illinois River basins starting in 2021. Similarly, the U.S. Forest Service has begun planning for a Forest Service Soil Moisture Monitoring Network in coordination with the NCSMMN. All the above activities, being conducted in coordination with or under the auspices of the NCSMMN, will serve to extend and improve soil moisture monitoring across the United States and support nationally relevant product development.
Future uses of the NCSMMN would include inclusion in the decision making for the National Drought Monitor in the United States to help improve the accuracy of drought estimates. Improved satellite calibration and validation of model and satellite products would also be possible. Numerous decision support systems related to agriculture, forestry, and hydrology will benefit with an improved network of realtime in situ measurements to quantify one of the most critical parameters at the land surface atmosphere interface.
In conclusion, there must be a strategic and coordinated effort to utilize and expand in situ soil moisture monitoring across the United States. The NCSMMN will coordinate this process. The collection of high-quality soil moisture data can be a complicated and challenging process, but it is ultimately necessary to coordinate disparate networks, if the value of soil moisture data is to be fully realized and connections between broader agencies and applications can demonstrate the value of soil moisture resources.

A C K N O W L E D G M E N T S
More than 50 individuals from a wide range of organizations, including federal, state, and local agencies, academia, the private sector, and nongovernmental organizations worked to develop the NCSMMN Strategy. We also thank B. Pellerin (USGS), K. Allander (USGS), and three anonymous reviewers for their comments on this manuscript. This research was supported by the U.S. Department of Agriculture, Agricultural Research Service. USDA is an equal opportunity employer and provider.