Curating Quantitative Methods to Detect Long-Term Trends in Phenological Species Monitoring

Abstract

Phenological species monitoring is crucial to determining the impacts of climate change on key indicator species (Diez et al. 2012, Ibáñez 2010). The Old Woman Creek National Estuarine Research Reserve (OWC NERR) is one of two reserves operating under the National Estuarine Research Reserve System (NERRS): a division of the Ohio Department of Natural Resources (ODNR) established in 1960 (Herdendorf et al. 2006). The OWC NERR runs a phenological program with 9 species monitoring initiatives including all-inclusive avian species, nest box monitoring with native cavity nesters such as tree swallows, bald eagle nesting data, and lungless salamander population data. The OWC NERR conducts research through citizen science by having volunteers make observations and collect data across initiatives over time. Using citizen science as a collection method may lead to the introduction of specific forms of biases and errors that must be overcome to draw meaningful conclusions. The OWC NERR asked our team to examine data from the selected initiatives and to attempt to answer the research questions they provided. To do this we made a Quality Assurance Quality Control plan for data collection and data entry for recommended adoption into OWC NERR’s protocols and to make data clean-up for us and future research teams more manageable. Data clean-up consisted of a master R validation script that was adapted for the selected initiatives. We ran into common data entry errors (i.e., misspellings, inconsistencies, different name entries for the same observer) and flagged missing data or data that fell outside of the program’s protocols. After the data clean-up, we created data analysis scripts for each selected initiative, all ran an “na_frac” table to determine what made up the missing data, most of which was weather entry data. We all set about answering the research questions provided through modeling and graphical analysis as well as determining the number of biases and errors there may be to continually keep new data entries up to protocol requirements without having to constantly change the clean-up validation script. After the analysis, we compiled a list of recommendations to present to OWC NERR to help develop a better Quality Assurance Quality Control protocol in which data collection and entry is consistent and to remove certain biases and errors. We also included recommendations for the individual initiative protocols and recommended to remove the need for volunteers to enter weather data and instead use the on-site weather station and water quality SWMP data collection that also runs out of OWC NERR. Also, to include detailed pictures or offer the use of an app like Canapeo to determine exact cloud cover and vegetation cover. Finally, we recommend overall that each initiative should implement key life stage observations to make it easier to go through the data for specific information that can lead to better educational material and volunteer training.