Challenges with secondary use of multi-source water-quality data in the United States

Abstract

Combining water-quality data from multiple sources can help counterbalance diminishing resources for stream monitoring in the United States and lead to important regional and national insights that would not otherwise be possible. Individual monitoring organizations understand their own data very well, but issues can arise when their data are combined with data from other organizations that have used different methods for reporting the same common metadata elements. Such use of multi-source data is termed "secondary use"-the use of data beyond the original intent determined by the organization that collected the data. In this study, we surveyed more than 25 million nutrient records collected by 488 organizations in the United States since 1899 to identify major inconsistencies in metadata elements that limit the secondary use of multi-source data. Nearly 14.5 million of these records had missing or ambiguous information for one or more key metadata elements, including (in decreasing order of records affected) sample fraction, chemical form, parameter name, units of measurement, precise numerical value, and remark codes. As a result, metadata harmonization to make secondary use of these multi-source data will be time consuming, expensive, and inexact. Different data users may make different assumptions about the same ambiguous data, potentially resulting in different conclusions about important environmental issues. The value of these ambiguous data is estimated at $US12 billion, a substantial collective investment by water-resource organizations in the United States. By comparison, the value of unambiguous data is estimated at $US8.2 billion. The ambiguous data could be preserved for uses beyond the original intent by developing and implementing standardized metadata practices for future and legacy water-quality data throughout the United States.

Unexpected stasis in a changing world: Lake nutrient and chlorophyll trends since 1990

“…Nutrient, chlorophyll, and lake depth data came from LAGOS-NE LIMNO (Version 1.054.1), a database that integrated field-based measurements of nutrients and physical properties from 54 government and university organizations on a subset of lakes across the study extent. Additionally, because a major focus in creating LAGOS-NE was in compiling and integrating metadata for each dataset, we overcame a common problem that is found with simply downloading data from a government repository (e.g., the U.S. Environmental Protection Agency's STOrage and RETrieval Data Warehouse)-namely that as much as 50% of the data values can lack sufficient metadata to use the data effectively (Sprague, Oelsner, & Argue, 2017). LAGOS-NE GEO is a collection of spatially referenced measurements that provides contextual information for all lakes >4 ha in the study extent.…”

“…Because LAGOS-NE includes a census of all lakes >4 ha in the study extent (see additional file 9 in Soranno et al, 2015 for a description of lake inventory), we were able to assess how the ecological context of lakes with chemistry data (hereafter, "sample lakes") compared to the population of lakes in the region (hereafter, "census"). Additionally, because a major focus in creating LAGOS-NE was in compiling and integrating metadata for each dataset, we overcame a common problem that is found with simply downloading data from a government repository (e.g., the U.S. Environmental Protection Agency's STOrage and RETrieval Data Warehouse)-namely that as much as 50% of the data values can lack sufficient metadata to use the data effectively (Sprague, Oelsner, & Argue, 2017). A detailed description of how LAGOS-NE was built, including details on data sources and methods of metric derivation, can be found in Soranno et al (2015).…”

Scientific integrity issues in Environmental Toxicology and Chemistry: Improving research reproducibility, credibility, and transparency

“…Unlike method development, however, the reporting and interpretation of environmental chemistry has common pitfalls, particularly in analyses from large data sets or compiled databases, and in citation practices. For example, metadata specifying fundamental details may be missing or misunderstood, such as whether concentrations of metals or other elements in water are from filtered or unfiltered samples or whether they reflect the total mass of the element or only 1 speciation state (Sprague et al ). Aquatic metals concentrations declined from milligram‐per‐liter levels in reports from the 1980s to microgram‐per‐liter or submicrogram‐per‐liter levels by the late 1990s.…”

Do we know how much fluvial sediment reaches the sea? Decreased river monitoring of U.S. coastal rivers

“…Groten & Johnson, ; Walling, Webb, & Woodward, ). Additionally, the techniques and other essential metadata of these records may not have been properly documented in many of the original data sources, such that their utility for developing estimates of fluxes or long‐term databases for analyses of change may be considered suspect (Sprague, Oelsner & Argue, ). Combined, this suggests that there may be value in compiling data from these broad efforts, but it is still unclear how these different data can be best used for estimating sediment fluxes.…”