RCUE Aerial Algae Monitoring
- Carol Thornber (URI Biological Sciences)
- Chris Deacutis (RI Department of Environmental Management)
- Giancarlo Cicchetti (US EPA).
- Jordan Kirby
- William Fanning
Harmful macroalgal blooms occur in coastal estuarine systems worldwide. Algal blooms (dense, floating aggregations) can create significant ecological and economic effects on coastal communities. Bloom decomposition can lead to hypoxia-induced fish kills, create noxious smells, impact fisheries, and detract visits by tourists and locals. Marine fish distributions in Narragansett Bay have already shifted due to climate change, and many other species, including bloom-forming macroalgae, are predicted to increase in density and distribution due to increased growth rates from higher water temperatures.
Developing innovative, rapid methods of monitoring coastal systems for these blooms is of critical importance, as current techniques involve either time-intensive quantitative surveys or expensive aerial manned helicopter surveys. The current state of the art for macroalgal monitoring relies on painstaking manual quadrat/transect surveys (Figure 1). These quantitative limited-area assessments can only be conducted during low tides. While PI Thornber has a long-term dataset (started in 2005; e.g. Guidone et al 2013 Harmful Algae), the combination of labor intensive sampling with time pressure severely limits the geographic coverage of the data collected. PIs Thornber and Deacutis have also previously used aerial surveys with manned helicopters to gather data on macroalgal blooms for several years. However, operating a helicopter is expensive, and many areas with high algal abundance cannot be properly surveyed because of flight path restrictions to the PVD airport. Collection of high quality imagery is difficult without special camera mounts, pilots trained in research methods, and helicopters with specialized navigation equipment, all of which are costly.
Starting in March 2014, we will work to demonstrate a novel and cost effective technique -- aerial imaging of macroalgal blooms by small autonomous robotic helicopters -- which can be used to measure changes in ecosystem structure over time. To validate the technique, we will perform robotic surveys ground-truthed by simultaneous quantitative biomass collection surveys. We will demonstrate that (a) an aerial robot can provide complete image coverage of a large sample area; (b) we can properly associate the resulting images with sampling locations on the ground; and (c) we can accurately estimate macroalgal blooms percent cover, biomass, and dominant species composition from the aerial images combined with limited on the ground sampling.
Using an image processing technique known as Structure from Motion (SfM) the aerial images can be used to create a highly accurate geo-referenced 3-D surface. The images above show Round Hill Beach, in Massachusetts in a mosaic produce using AgiSoft on images from a 3DRobotics IRIS+ quadrotor.