Innovation in Restoration: Estimating Seed Counts Using a Photography App

Samantha Kelly, Cessair McKinney, and Dr. Kerry Byrne, Environmental Science and Management, Humboldt State University, 1 Harpst St, Arcata, CA 95521

The field of restoration ecology is growing rapidly, and scientists continue to search for innovative techniques to make restoring degraded ecosystems more effective. Restoration projects involve restoring plant communities through direct seeding. Seeds are often acquired from nearby reference communities, as there is some evidence that local seed may be more adapted to prevailing conditions. It can be difficult to quantify the number of seeds collected, particularly for species with small or variable seed size. The process of accurately quantifying seeds is labor-intensive and therefore costly. A rapid alternative to manually counting seed might be to estimate seed quantity through a photo counting software. The objective of our study was to assess the accuracy and reliability of a user friendly photo counting software to estimate the number of seeds for two small-seeded sagebrush species; Artemisia cana and Artemisia arbuscula. We analyzed the accuracy and reliability of the photography application (app) “CountThings” to count the seeds of these two species of sagebrush. First, we manually counted seed samples then compared these values to the estimate calculated by the photography app. We conducted a linear regression to compare the estimated seed counts to the actual seed counts. We manually counted a total of 34 seed samples, ranging in size from 7 to 438 seeds with an average of about 183 seeds. We found that the number of seeds estimated  using “CountThings” was similar to the numbers of seeds manually counted (R2 = 0.92). The incorporation of photography software to count seeds for restoration is uncommon and there is little scientific literature on the subject beyond seed counting for agricultural purposes. The accuracy of “CountThings” suggests that photo counting apps could be a useful tool for ecologists and land managers who desire a rapid method to accurately estimate seed numbers for restoration purposes. 

Additional Abstract Information

Presenters: Samantha Kelly , Cessair McKinney

Institution: Humboldt State University

Type: Oral

Subject: Ecology

Status: Approved

Time and Location

Session: Oral 2
Date/Time: Mon 3:00pm-4:00pm
Session Number: 216
List other presenters in this same room and session