Abstract Detail

Biodiversity Informatics & Herbarium Digitization

Haughland, Diane L. [1], Crisfield, Varina [1], Huggard, David [2], Dennett, Jacqueline [3], Nielsen, Scott [3].

Long live the census: the (floristic) census is dead.

Many studies have shown that imperfect detection significantly affects floristic surveys, even when conducted by expert botanists. Detectability, the probability of recording a given plant, can vary with plant-related traits such as growth form, phenology, and community composition, as well as with observer- and survey-related parameters such as training, experience, plot size, and survey duration. Given these sources of error, we asked how we could optimize the accuracy and repeatability of vascular plant surveys conducted by novice to intermediate technicians working for a large-scale biodiversity monitoring program in Alberta, Canada. In a series of field experiments in the grassland and forested regions of Alberta, we explored how time limits, plot size, teamwork, and target species affected survey results. Each set of experiments involved 6-8 technicians and 2 expert botanists to provide quality control and best-case scenario comparisons. Our results show that even in small plots (1m²), working in teams, and with unlimited time to record and collect plants, novice to intermediate technicians typically could not attain the level of species capture recorded by the experts when asked to conduct a census-style survey. This disparity was amplified in more species-rich or complex vascular plant communities. Conversely, we observed significantly higher repeatability when we reduced the number of species the technicians were responsible for recording to a predefined list of approximately 130 species. Technicians also were provided training on those species, but resources did not allow us to definitively separate the effect of the targeted training from the effect of reducing the survey to the predefined species list. Our findings have implications for monitoring programs that employ large numbers of observers over time, as well as citizen science studies that rely on many observers with varying skill levels.

1 - Royal Alberta Museum, Alberta Biodiversity Monitoring Institute, 9810 103a Ave, Edmonton, AB, T5J 0G2, Canada
2 - University of Alberta, Alberta Biodiversity Monitoring Institute, CW 405 Biological Sciences, Edmonton, AB, T6G 2E9, Canada
3 - University of Alberta, Renewable Resources, 705 General Services Building, Edmonton, AB, T6G 2H1, Canada

Keywords:
plant survey
detectability
observer effects
survey design
Alberta Biodiversity Monitoring Institute
monitoring.

Presentation Type: Oral Paper
Session: BIHD1, Biodiversity Informatics & Herbarium Digitization
Location: Tucson I/Starr Pass
Date: Monday, July 29th, 2019
Time: 5:00 PM
Number: BIHD1014
Abstract ID:1048
Candidate for Awards:None