Oregon Ash in Tryon Creek Watershed

In the interest of gathering an inventory of Oregon Ash and monitoring for inevitable Emerald Ash Borer infestations in the Tryon Creek watershed, Tryon Creek Watershed Council is mobilizing trained volunteers to gather field observations of Oregon Ash and their conditions throughout the watershed starting in fall 2023. To support these efforts, I have created a suitability map identifying likely habitat for Oregon Ash in the watershed, as well as a form in Survey123 that users can fill out to document location and health indicators for Oregon Ash. The data from the form is georeferenced and could be visualized on the ArcGISOnline web mapping application, or downloaded as a .csv and imported into a GIS or other data visualization software, as I did for this analysis using ArcGIS Pro.

Oregon ash typically grows in lowland conditions and typically prefers flat or low-angle slopes. The slope calculations for the watershed were derived from a 10m DEM produced by the United States Geological Survey, obtained through the USGS National Map Downloader.

Soil is considered hydric when it " is saturated, flooded or ponded long enough during the growing season to develop anaerobic conditions in the upper part of the soil profile that favor the growth and regeneration of hydrophytic vegetation. ^"5  A soil is inundated when the water table is at or above the soil surface. A soil is flooded if the water is moving across the soil surface as in a slough or on a floodplain." Oregon ash is known to prefer "moist, boggy lowlands" and "is characteristic in seasonally flooded habitats". ⁶  This dataset was produced by the Wetland Conservancy and obtained via Oregon Spatial Data Explorer.

Oregon ash typically prefer wet soil conditions, and this variable accounts for the soil capacity to store water during periods of low or no precipitation. All soils data was derived from the Gridded National Soil Survey Geographic Database, produced by the USDA NRCS.

Oregon ash is most commonly found in "deep, poorly drained clays or silty clay loams that are rich in humus". However, it is also found in "sandy, rocky, and gravelly soils in riparian areas or areas with seasonal flooding" ^6.  The soil texture composition for Tryon Creek is depicted in this map. The watershed is predominantly silt loam, followed by silty clay loam. Closer to the river small pockets of gravelly clay, silty clay, and unweathered bedrock occur. ² 

Since ash prefer poorly drained soils, gNATSGO data was queried for a dominant drainage class "poorly drained" within Tryon Creek watershed. Most of the soils in the watershed are characterized by poor drainage.

Oregon ash occurrence in Tryon Creek watershed is unknown, and only three iNaturalist observations in the area are currently available, not enough to determine the relationship between suitability factors and species occurrence.

Without a background in statistics, and in the interest of avoiding arbitrary assignment of weights to variables, an alternative method was devised to identify statistical relationships between the variables and occurrence of Oregon ash. A DEM at 1/3 arc-second resolution (approx. 10m ² )with an extent of 1x1 degree of latitude (approx. 8500km ² ) surrounding the area of interest was selected as a boundary, and input variables were clipped or generated within this boundary and converted to raster. 693 Oregon ash observations captured with iNaturalist and occurring within the selected boundary were processed using the Extract Multi-Values to Point geoprocessing tool.

This processing tool finds the raster cell underlying a point observation and imports the value of that cell into the attribute table. So for each input layer, the value of the cell was added to the row assigned to each of the 693 point observations. This process was selected in order to aggregate the values of each input to the attribute table of the point observation feature class so that ArcGIS Pro can be used to generate statistics.

Variables best described in binary terms (e.g. "True": inside a wetland area, or "False": outside a wetland area) were ascribed a value given a Boolean condition (True = 10, False = 1) and weighted accordingly in the final model.

Hydric Soils contained 462/693 observations, or 66.67% of total observations in the selected area.

Wetlands buffered to 50 meters contained 440/693 observations, or 63.40% of observations in the selected area.

Poorly Drained Soils (with a dominant drainage class of either “somewhat poorly drained” or “very poorly drained”) accounted for 131/693, or 18.90% of observations in the selected area. These findings indicate that wetlands and hydric soils are likely to have a signficant impact on whether conditions are favorable, while poor soil drainage may be present but is likely less important.

The Reclassify geoprocessing tool was used to assign a range of values to each input raster on a 1-10 scale. The final suitability surface was then created using the Weighted Overlay geoprocessing tool, which allows the user to input the reclassified raster layers and weight their influence on the final output as a percentage. The parameters I chose were: Slope- 20 % Wetlands- 22% Wetland (Hydric) Soils -24% Poorly Drained Soils- 8% Water Storage- 14% Soil Texture-12%

This project represents an exploratory analysis of likely suitable habitat for Oregon Ash in the Tryon Creek Watershed, and a Survey123 form to gather data on their occurrence and condition. The model fit is constrained by a lack of available Oregon ash data within the watershed, and my own limitations in analyzing statistical relationships. However, from what I was able to glean reading descriptions from OSU Extension, USDA Forest Service, and Oregon Department of Forestry, as well as the statistical relationships derived from the DEM-extent observations, my hope is that this habitat suitability analysis will indicate with some accuracy the distribution of Oregon ash in the watershed. Further research could be done to fine-tune the model, including expanding the number of input variables, conducting more robust statistical analysis, and updating the model once a robust sample of field observations is collected within the watershed.

The Survey123 data will also provide insight into the number and distribution of Oregon ash in the watershed, as well as a range of conditions relevant to monitoring for EAB infestation. This data could also be useful in tracking EAB spread through the region and identifying areas where there is signifcant impact. It seems an unfortunate eventuality that restoration work will be required in many riparian ecosystems to replace and approximate the role of Oregon ash, and so it is also my hope that this analysis could be modified and updated for similar species to that end.

1. Oregon ash (Fraxinus latifolia) | Oregon Wood Innovation Center. (n.d.). https://owic.oregonstate.edu/oregon-ash-fraxinus-latifolia
2. Brissette, (2015, June 22). Oregon ash, Fraxinus latifolia. Native Plants PNW. http://nativeplantspnw.com/oregon-ash-fraxinus-latifolia/
3. Emerald ash borer beetle. USDA APHIS | Emerald Ash Borer Beetle. (n.d.). https://www.aphis.usda.gov/aphis/resources/pests-diseases/hungry-pests/the-threat/emerald-ash-borer/emerald-ash-borer-beetle
4. Oregon Department of Forestry. (n.d.). Forest Health Fact Sheet - Oregon.gov. Forest Facts: Emerald Ash Borer (EAB) Agrilus planipennis Fairmaire. https://www.oregon.gov/ODF/Documents/ForestBenefits/Woodboringbeetles.pdf
5. Florida Department of Environmental Protection. (n.d.). Wetland Delineation - Hydric Soils. Wetland Delineation- Hydric Soils. https://floridadep.gov/water/submerged-lands-environmental-resources-coordination/content/wetland-delineation-hydric-soils
6. Owston, P. W. (n.d.). Fraxinus latifolia Benth. Oregon Ash. https://www.srs.fs.usda.gov/pubs/misc/ag_654/volume_2/fraxinus/latifolia.htm

I would like to thank Alexis Barton from Tryon Creek Watershed Council for her guidance and encouragement on approaching this complex topic, and collaboration in creating the Survey123 form. Sean McKenzie from Oregon Department of Forestry for comparing notes on our project early on and presenting the idea to run a logistic regression to quantify the influence of input variables. And finally, my classmates Ian Maher, Valence Brenneis, and Ruth Bonnette for their feedback and assistance in resolving technical problems at key moments in this process. I would also like to thank Christina Friedle and PCC for providing me this opportunity.