SYMP 14-3 - Coupling citizen science with precipitation and phenology monitoring and modeling to support invasive species management

Wednesday, August 5, 2009: 2:20 PM
Blrm B, Albuquerque Convention Center
Aaryn D. Olsson1, Michael A. Crimmins2, Barron J. Orr1, Start E. Marsh3 and Willem J.D. van Leeuwen3, (1)Office of Arid Lands Studies, University of Arizona, Tucson, AZ, (2)Department of Soil, Water and Environmental Science, University of Arizona, Tucson, AZ, (3)Office of Arid Lands Studies, University of Arizona
Background/Question/Methods

Citizen science is emerging as method for expanding sensor networks while engaging the public in outreach activities. In the US Southwest, public engagement at all levels is critical to managing buffelgrass, an exotic invasive grass that threatens much of the Sonoran Desert with a devastating grass-fire cycle. Most native flora cannot withstand repeated fires, including the iconic saguaro cactus (Carnegiea gigantea). Unfortunately, the most effective means for control is limited to several weeks in the summer. Herbicide treatments require actively photosynthesizing buffelgrass, which varies over space and time due to heterogeneous rainfall patterns associated with convective activity during the summer monsoon. Green-up and senescence—the two most important phenological events managers are interested in—are triggered primarily by soil moisture during the summer when temperatures are not limiting, and soil moisture fluctuates as soil moisture rapidly dissipates due to evaporation and transpiration. We present a decision-support system based on citizen volunteered rainfall measurements and buffelgrass phenology observations to direct region-wide strategic efforts to fight buffelgrass in real-time. A soil moisture model derived from interpolated rainfall measurements made by volunteers in the Rainlog network (http://rainlog.org) is used to predict buffelgrass green-up and senescence. Evaporation is modeled using the Thornthwaite equation. Buffelgrass phenology will be validated with ground-based observations of additional volunteers.A cross-jurisdictional buffelgrass control team will be dispatched across Tucson, AZ, and surrounding areas to treat high-priority infestations when treatments will be most successful.

Results/Conclusions

We used 100 days of precipitation data from 2008 to evaluate interpolation techniques and assess data quality. Preliminary results suggest that spatial regression can be used to pre-process the precipitation data to identify outliers for exclusion from the spatial interpolation. We are currently developing the phenological model and training citizen scientists to map buffelgrass and recognize its phenological stages. The model will start being used operationally by the Southern Arizona Buffelgrass Coordination Center in the summer of 2009 before the onset of the summer monsoon.

Copyright © . All rights reserved.
Banner photo by Flickr user greg westfall.