Using these numbers, we can roughly estimate the money spent on research for which the data are unrecoverable (70% of datasets): approximately $105 million USD was spent collecting data that are no longer recoverable and, therefore, effectively non-existent to science. Of the 419 projects funded by EVOSTC during this time, only 27% of the datasets were able to be recovered after a total of 5 years hunting down datasets, this grew to 30%. įrom 2012 to 2014, a group of researchers from the National Center for Ecological Analysis and Synthesis worked to recover the historical datasets funded by EVOSTC, focusing specifically on data collected between 19. In their annual report from 2010, the EVOSTC notes that some $151.2 million USD were spent on “research, monitoring and general restoration” during 1992–2010 fiscal years. Prior to 2003, there was no requirement for data preservation or availability afterwards, all projects were awarded under explicit conditions from EVOSTC that data be preserved and made publicly available. A large portion of funds were directed towards determining and monitoring the impacts of the oil spill on oceanographic, environmental and ecological conditions. The Exxon Valdez Oil Spill Trustee Council (EVOSTC) was established in 1991 to oversee the spending of funds from a civil settlement in 1991 between Exxon, the United States federal government and the state government of Alaska. As a result, approximately 37 000 tonnes of crude oil spilled into the sound, leading to catastrophic short- and long-term ecological consequences.

In 1989, the oil tanker Exxon Valdez struck the Bligh Reef in Prince William Sound, less than 2.5 km from the Alaskan shore. The ultimate goal of data rescue is to make previously inaccessible or poorly preserved data available for (re)use, ideally through archiving them in a permanent, publicly accessible and reusable format.īox 1. Data rescue-defined here as the identification, preservation and sharing of valuable data and associated metadata at risk of loss-has the potential to realize substantial benefits for society, especially considering the crucial roles that baseline data play in informing management and policy decisions. Languishing data, however, also offer an enormous opportunity. In their unusable and 'at-risk' state, these data represent an egregious waste of resources expended on their collection ( box 1). Yet, despite their substantial value, data are often misplaced, filed away or otherwise rendered unusable, often through poor data management practices. aggregation, collation and synthesis) of data from different contexts is essential to establishing broader ecological knowledge and informing conservation management. While data collection is often targeted to particular populations, communities or locations, the reuse (i.e.

Observational and experimental data derived from ecology, evolution, conservation, and environmental sciences (hereafter, environmental data) are essential to establishing historical trajectories of ecosystems (baselines), understanding how species and communities respond to environmental change and designing and evaluating the outcomes of management efforts. It is, therefore, imperative that we identify and preserve valuable, at-risk environmental data before they are lost to science.ĭata are among the most valuable outputs of research and scholarship beyond helping answer important questions, they inform new lines of inquiry, new testable hypotheses and future data collection efforts. In an era of rapid environmental change, the best policy solutions will require evidence from both contemporary and historical sources. We discuss prioritization of datasets to rescue, forming effective data rescue teams, preparing the data and associated metadata, and archiving and sharing the rescued materials. Here, we outline seven key guidelines for effective rescue of historically collected and unmanaged datasets.

data reuse), and lacks general recommendations. While rescuing data is not new, the term lacks formal definition, is often conflated with other terms (i.e.

Improvements in policies and best practices around data management will hopefully limit future need for data rescue these changes, however, do not apply retroactively. The practice of data rescue, which we define as identifying, preserving, and sharing valuable data and associated metadata at risk of loss, is an important means of ensuring the long-term viability and accessibility of such datasets. Although immensely valuable, these data are at risk of being lost unless actively curated and archived in data repositories. Historical and long-term environmental datasets are imperative to understanding how natural systems respond to our changing world.

Data rescue 2