4

Lessons Learned from the Listening Sessions

In a brief session between the day’s two panel discussions, workshop organizing committee member Alison Marklein from the University of California, Riverside, described the organizing committee’s learning from 1 year of talking with different organizations that collect or curate soil data. National Academies organizing committees usually do not conduct listening sessions, but the almost year-long delay in holding the workshop caused by the COVID-19 pandemic presented an opportunity to meet with different organizations in an effort to understand the state of dynamic soil information systems.

Over the past year, Marklein said, members of the organizing committee met with 17 different organizations, including U.S. agencies, international agencies, and members of the private sector (see Table 4-1). In these meetings, the organizing committee heard several common themes. First, the potential for increased inter-agency communication and collaboration in this area is high. Second, there are few dynamic soil data. Third, funding is precarious for efforts to garner a long-term dynamic understanding of soils.

The organizing committee members posed the following questions to each organization:

• What is your vision, and what do you want to do with the soil data?
• What is working well with your current database or data collecting effort?
• What are the roadblocks?
• How are data curated, transferred, analyzed, and shared?
• What are the drivers of change?
• What do you want to be able to do?
• How are your data being used and who uses them?
• What infrastructure is needed to capture and store the data?
• At what spatial and temporal scales are different variables measured?

The organizing committee uncovered several U.S.-based and global data products. The Natural Resources Conservation Service operates the National Soil Information System as well as the Soil Survey Geographic Database and a closely related database, the Gridded

TABLE 4-1 Organizations Listened to by the Workshop Organizing Committee

NOTE: CSIRO = Commonwealth Scientific and Industrial Research Organisation; DataONE = Data Observation Network for Earth; iSDA = Innovative Solutions for Decision Agriculture; ISRaD = International Soil Radiocarbon Database; NEON = National Ecological Observatory Network; NOAA = National Oceanic and Atmospheric Administration; USDA = U.S. Department of Agriculture.

SOURCE: Marklein, slide 3.

National Soil Survey Geographic Database. The National Oceanic and Atmospheric Administration (NOAA) has the National Integrated Drought Information System and the National Coordinated Soil Moisture Monitoring Network. The Soil Health Partnership has a database, the National Ecological Observatory Network has 15 active soil data products, and the U.S. Forest Service operates the Forest Inventory Analysis. Global data products are available from the Data Observation Network for Earth, the Global Soil Information System, the International Soil Radiocarbon Database (ISRaD) (see more discussion on ISRaD in Chapter 2), the Global Soil Partnership, and the International Soil Reference and Information Centre.

From its listening sessions, the organizing committee noted several common challenges facing soil-related organizations. One significant challenge is obtaining continuous funding for monitoring. “A lot of the agencies are really excited about funding science-driven ques-

tions that are a few years in scope,” Marklein said, “whereas we really need to have this long-term monitoring … to get long-term effects.” A second challenge is that many of the organizations are understaffed in the areas of soil science and data analytics. Other challenges relate to data privacy and security and harmonization of data naming conventions; common methodologies, including sampling and analysis procedures; errors associated with methodology, analysis, and facilities; capture of spatial and temporal data at varying scales; and re-sampling of destructive samples.

Marklein described two organizations in more detail to illustrate the types of data being collected and curated by different groups and the types of challenges they face. First, the European Commission Joint Research Centre aims to provide scientific evidence and data to support policy making about soil-related issues, including agriculture, the environment and climate change, biodiversity, and human health. The soil information system has data from 27 countries that have been harmonized and are free to access. The data report physical and chemical properties and soil biodiversity in the form of DNA sequencing, and there is some information on land use type. The spatial resolution is roughly 2 km², but sampling is done on roughly a 14-by-14-km grid. The temporal resolution is every 3 years, and the same sites are resampled for land use and land cover change. The database is regularly updated as new data become available.

To address the challenge of inconsistent data analysis across laboratories, the center charged one laboratory with performing all of the analyses. However, this solution introduced another challenge—reporting delays because of the bottleneck from having only one laboratory do analysis. Furthermore, spectral and remote sensing data involve a lot of uncertainty, and interest and capabilities in these data vary among the European Union’s member states, some of which do not have monitoring systems. Another major challenge is data privacy because sampling sites are often on private lands and, in particular, data on the presence of heavy metals is considered to be private.

In terms of success, the center has achieved a harmonized dataset for soils for Europe. The data are used by different stakeholders, including scientists, modelers, policy makers, and farmers, and strong collaborations exist among organizations in that space.

The second example presented by Marklein is NOAA, which has the National Coordinated Soil Moisture Monitoring Network. The network’s goal is to collect multi-platform soil moisture measurements and merge in situ measurements and remote sensing data with numerical model outputs. NOAA partners with several federal and state agencies in the network, which has more than 150 end users. It operates 21 mesonets (meso-scale networks), which are networks of monitoring stations created to observe meso-scale meteorological and environmental phenomena. The main output of the network is measurements of soil moisture at 4-km resolution that are offered in near real time and regularly updated.

One of the network’s main challenges is how to best represent soil moisture. Should it be done in percentiles? In terms of anomalies? Via the U.S. drought monitoring categories, volumetric water content, or millimeters? A related challenge is how to communicate uncertainty for each of the ways to represent soil moisture. In addition, the underlying data maps are not currently available because of funding issues. Finally, there are a variety of measurement and data challenges, for example, how to recognize the uncertainty and validity in sensors, how to best integrate data given spatial distribution issues and questions about the representativeness of each point, and how to deal with data gaps.

The soil monitoring network has been operational for more than 1 year. Marklein presented an example of the sorts of soil moisture maps that it produces, in this case the Blended Soil Moisture Product (5cm All Blend), which is a high-resolution, gridded soil moisture map derived from in situ, model-generated, and satellite data.

The network is in the process of developing a cyber infrastructure, Marklein said, and its next steps include interpolating soil moisture data in space, time, and depth; blending different data sources; furthering validation and quality control; and, if funding becomes available, making the underlying data accessible. A longer-term goal is to project soil moisture.

Marklein also displayed charts that compare the physical, chemical, and biological properties of soil that eight different organizations collect. This comparison demonstrates that parameters vary among soil data collection efforts.

Marklein summarized some common needs that emerged from the listening sessions:

• funding for monitoring for long-term understanding
• more repeated measurements in the same locations
• more communication between agencies
• increased clarity on nomenclature
• more metadata on sampling methodology and processing
• data at a scale relevant to farmers
• archived soils for future analysis

During the discussion session that followed Marklein’s presentation, Tomislav Hengl from the OpenGeoHub Foundation added the foundation to the list of global data producers, pointing participants to the resources available at its website (see Continuing Engagement Opportunities in Chapter 8). The hub is a transparent system, and Hengl and colleagues are working to convert it into a community system so that people can pull, modify, add, and contribute to the data.

Michael Young from The University of Texas at Austin noted the three issues related to data standardization: standardization of data collection, standardization of data storage, and harmonization of data so that their different formats, scales, and types can be combined. He asked Marklein whether the organizations discussed the sticking points related to each of these issues. Marklein responded that organization representatives expressed a strong interest in increased harmonization of datasets, but standardization is difficult to achieve, especially when not everybody is in the room at the same time. One goal of this workshop, she added, is to start this conversation so that people not only learn about what other data are available, but also identify ways to make the datasets interoperable.