Joint Pilot Fish Habitat Framework
Integrating Tidal and Non-Tidal Waters in the Patuxent River Basin
Start with a review of previous fish habitat assessment methodologies and results
Explore the relationships of various summary frameworks including: 2-D river reach and catchment framework for inland rivers and streams, hexagonal, square, or voxel grids for estuarine habitat areas, and/or hybrid, multi-scale grids for common application for headwaters to estuary summaries
It was determined that neither the hex grid nor line/polygon method was particularly suitable for an inland to tidal habitat assessment so a square gridded framework was selected. We will talk more about this in the Gridded Framework section of this StoryMap.
Next, develop criteria for selecting a tributary of focus, use criteria to propose a tributary, and seek input from stakeholders
We conferred with and sought approval from the FHAT on the tributary river basin we should use for testing. Potential rivers selected needed to include: Cold Headwater, Large Non-tidal Rivers, Tidal Freshwater, and Tidal Estuarine. Based on the presence of these four habitat types, the size of the tributary, and the management complexity (i.e. occurring within a single state's jurisdiction), the Patuxent, Rappahannock, and James Rivers were chosen as prime candidates for pilot framework development (see below) while the other river systems were excluded.
RIVER SYSTEMS WITH ALL 4 HABITATS AND 1- 2 FISHERY MANAGEMENT JURISDICTIONS
| RIVER SYSTEMS WITH 3 HABITATS AND 1- 2 FISHERY MANAGEMENT JURISDICTIONS
|
|---|---|
RIVER SYSTEM WITH ALL 4 HABITATS AND MULTIPLE JURISDICTIONS
| RIVER SYSTEM WITH 3 HABITATS AND MULTIPLE JURISDICTIONS
|
River System Groups Based on Criteria. River Basins that are crossed out were considered for the joint pilot but ultimately not chosen.
Next, coordinate efforts on stakeholder engagement with interest groups to determine species and issues of interest as it relates to fish habitat assessment
The team sought input from members of the Patuxent River Basin Commission , the Maryland Department of Planning , and local residents (via outreach events with the University of Maryland Integration and Application Network ( ian.umces.edu )).
While much of the feedback from stakeholders was varied in focus and scope, they shared the following elements...
Click on the right arrow to scroll through the feedback
Next, collate and organize fish data
The team collated fish datasets for selected species of interest that contrasted life history and habitat use characteristics, including headwater resident fish (tessellated darter), estuarine resident fish (white perch), and diadromous fish that move between headwaters and the estuary (American eel). Due to temporal, spatial, and methodological differences between sampling studies used to obtain species data, it was difficult to identify true absences within the entire basin extent. Therefore we calculated pseudo-absences to use during the modeling process using a statistical approach.
Maps below show species presence locations
Next, construct the analysis framework
The team developed a gridded analysis framework using the USGS Coastal National Elevation Database (CoNED) as a base. This dataset is an integration of the best available bathymetric data (sonar, soundings) and high resolution topography (from lidar), and is available for selected coastal regions in the United States, including the Chesapeake Bay and Delaware Bay. We chose this dataset as the basis for our framework as it seamlessly integrated upland topography with bathymetry, allowing flexible data summary techniques incorporating upstream flow and runoff influences as well as bottom characterization.
While the original resolution of the CoNED data is a 1m raster, we ultimately decided to assess habitat relationships at three spatial scales - square raster cells with sides measuring 1000m, 100m, and 10m.
Next, organize, summarize, and/or generate appropriate landscape and hydrologic predictors, including time-sequenced data
The team attempted to collect the most recent environmental data available, though various data sources extended over the last 30 years. Because many variables were not collected consistently over that time period, we choose to “flatten” the data over time and conduct the analysis as a snapshot in time rather than an assessment of change in habitat over time.
DEM Based Variables
Digital Elevation Model (DEM) data came from the USGS Coastal National Elevation Data product . DEM based metrics were used to identify heterogeneity and homogeneity of the landscape by computing terrain indexes, roughness, and slope. These DEM metrics were used to create flow accumulation metrics helping identify watersheds and flow paths based on topography.
Land Use/Land Cover
The team used land use and land cover (LULC) data from the 2013-2014 Chesapeake Conservancy ( Chesapeake Bay Program Land Use/Land Cover Data Project ) to better match the species occurrence data. LULC metrics were computed as the percentage of the LULC type within the cell (i.e. 40% of a grid cell was cropland). In addition to percent land cover maps, using the derived DEM metrics, we also included several metrics that described the aggregate LULC influence of the watershed respective to each grid cell.
Climate Data
Climate data were obtained from the PRISM Climate Group at Oregon State University ( PRISM Climate Data ) as monthly 30-year normalized datasets for air temperature minimum and maximum, and precipitation at an 800m resolution. Bioclimatic variables from the PRISM dataset were averaged on both monthly and quarterly scales to identify general climatic trends over the region.
Surface Water Quality
Water quality conditions were calculated using data from the Maryland Department of Natural Resources' “ Eyes on the Bay ” monitoring program and the Chesapeake Bay Interpolator. Summer (June-August) surface estimates were used for Dissolved Oxygen (DO), Water Temperature, and Salinity.
Substrate Bottom Type
Benthic habitat data was aggregated and curated by NOAA’s Chesapeake Bay Office and standardized to the Coastal and Marine Ecological Classification Standard .
Submerged Aquatic Vegetation
Distance to SAV beds was assessed using data collected and distributed by the Virginia Institute of Marine Science ( SAV Reports & Data | William & Mary ).
Hardened Shoreline
Distance to hardened shoreline was conducted using the most recent GIS shoreline inventory data release for each county within the tidal portion of the Patuxent watershed; hosted and collected by Virginia Institute of Marine Science ( Shoreline & Tidal Marsh Inventory | William & Mary ).
Protected Areas
Proximity to protected areas was utilized for this assessment, and the relevant information for the state of Maryland was downloaded from the Protected Area Database of the United States (PAD-US) database .
Benthic IBI
Data for benthic monitoring sampling events, sample collection, and Benthic Index of Biotic Integrity (B-IBI) collected by VERSAR were downloaded from the Chesapeake Bay Benthic Monitoring Program website. The average B-IBI score served as a key metric at each site, providing a comprehensive assessment of benthic health in that location.
Lastly, develop species and habitat distribution modeling applications using an ensemble of multiple model methods to predict probability of suitable habitat
Nested Ensemble Species Distribution Model (NESDM)
Initial results following our statistical analysis and modeling can be found in our NOAA Technical Memorandum . The statistical testing that we conducted was intended only to test the framework and modeling approach, and not to definitively predict all habitats where specific fish species might be present.
