Developing Innovative Tools for Managing Invasive Aquatic and Riparian Species.

The introduction and spread of invasive plant species poses a significant challenge to conservation practitioners and agency personnel responsible for managing native biodiversity. The development of innovative tools that bridge science and conservation practice to delineate target invasives will encourage a more proactive approach to their management and mitigation. This research is currently using hyperspectral remote sensing techniques to map aquatic and riparian invasive plants in the Sacramento-San Joaquin Delta (including the Tuolumne, Merced, and San Joaquin Rivers in particular) in the Central Valley of California. The geographic extent of the Sacramento-San Joaquin Delta (~3500 km²) makes the use of remote sensing techniques ideal for providing timely information on the spread of invasive species. The target invasive aquatic weeds are water hyacinth (Eichhornia crassipes) and Brazilian waterweed (Egeria densa) and the riparian pepperweed (Lepidium latifolium) and purple loosestrife (Lythrum salicaria). These aquatic and riparian invasive weeds are spreading at a phenomenal rate and incur enormous ecological and economic costs. There is a great need to provide accurate maps of the spatial distribution of key species to management and control agencies. In this study we are investigating the use of hyperspectral (Hymap) imagery which possesses both high spatial resolution (3m pixels) and high spectral resolution (126 bands) to identify riparian weeds, and both submerged and emerged aquatic weeds.

In 2003/2004, the University of California Davis, Calspace Center of Excellence, the California Department of Boating and Waterways (CDBW), and the California Department of Food and Agriculture (CDFA) cooperated on a pilot project to map the aquatic weeds water hyacinth and Brazilian waterweed, and the riparian pepperweed in the Sacramento San Joaquin Delta. Ten flightlines covering approximately 400 km² were acquired in June of 2003. There were sufficient differences in the spectral signature of the invasive plants to allow us to distinguish them from native vegetation. Data were analyzed using spectral mixture analysis (SMA) for the aquatic weeds and mixture tuned matched filtering (MTMF) for the riparian pepperweed. SMA and MTMF successfully mapped the distribution of the invasive aquatic and riparian species. Results show the target weeds were mapped with a classification accuracy of 90.6% when compared to 2003 sample sites. Acreage estimates for water hyacinth and Brazilian egeria were produced for the study site.

In 2004/2005 we are building upon the methodology developed in the pilot project and extending it to the entire Sacramento-San Joaquin Delta Region to provide the CDBW and CDFA with a comprehensive baseline for documenting the distribution and extent of invasive species in the entire Sacramento-San Joaquin Delta Region (Figure 1) as well as a rapid, cost effective, and valuable technique to make informed decisions relevant to the extent and spread of invasive species. We will continue to target the invasive aquatic weeds water hyacinth (Eichhornia crassipes) and Brazilian waterweed (Egeria densa) as well as the riparian pepperweed (Lepidium latifolium) and purple loosestrife (Lythrum salicaria). Hydrilla (Hydrilla verticillata) has not been found in the Delta in recent years, but has caused significant problems in other waterways throughout California. We will monitor the Delta for any new infestations of this aggressive aquatic weed.

In the long-term these results can be used in conjunction with other data, for example, surrounding land-use practices to develop correlations with the key species and to provide a basis for predicting their future distribution. The outcome of this research will be an increased understanding of the capacity and practicalities of image-based detection of invasive species and an assessment of the potential of predictive models to project their trajectories. Ultimately, this will provide state agencies and conservation practitioners with broadly applicable, science-based tools to better meet the challenges of controlling and monitoring invasive species.