In July and September 2012 the NERC airborne research and survey facility (http://arsf.nerc.ac.uk) collected data over the BESS-F3UES sites in Milton-Keynes, Luton and Bedford (see October 2012 news). The airborne remote sensing comprised aerial photographs, hyperspectral imaging data, showing visible and infrared reflectance in fine spectral resolution, and lidar data. These data are being used to characterise urban vegetation in unprecedented detail.
The hyperspectral imaging data can be used to map vegetation distribution across the urban zone and can help map species and vegetation status (as the health of each plant is indicated by their biophysical response to light in the visible and near-infra-red portions of the spectrum). The lidar uses a laser to measure 3D landscape structure precisely, allowing topographic maps and the height of objects above ground to be determined. Traditionally, this can be used to make a land cover classification map of tall vegetation (trees), short vegetation (grass and bushes), tall non-vegetation (buildings) and short non-vegetation (road or bare-soil).
The derived land cover maps provide a detailed model that can be used to address a range of scientific questions. For example, we are using these data to understand how birds travel through urban areas. However real urban areas are much more complex than these four classes. In the BESS-F3UES project we are taking the analysis a step further and using the “full-waveform” return of the lidar. This carries information about within vegetation structure, not just how tall a plant is but how dense the vegetation is at different heights above ground. The extra details we can determine from the full-waveform data are of great importance in guiding wildlife movements through urban areas – for example, a short shrub offers very different ecological resources to a bird compared to a well manicured lawn. Describing the complex 3D structure of the urban environment is the focus of the BESS-F3UES remote sensing group, based at the University of Exeter’s Penryn campus.
Developing new methods to measure vegetation from airborne remote sensing requires accurate validation or “ground truth” data. Traditional techniques for validating our aerial data would usually require time-consuming manual measurements of vegetation canopy characteristics by hand, either by crane or destructively sampling the vegetation canopy. The latter would not be feasible or popular with urban gardens. To address this need we have collaborated with Dr. Mat Disney of University College London (https://twitter.com/mathiasdisney, http://disneytls.blogspot.co.uk/) who spent two days with us using his terrestrial laser scanner (TLS) to measure gardens, parks and woods around Luton. TLS provides very accurate structural measurements of vegetation canopies, allowing us to cross compare with what the less detailed but wider coverage airborne lidar measures from above.
We are working on processing the full-waveform data. Our intention is to produce an urban habitat map with unprecedented detail and accuracy to provide insights into the function and flows of ecosystem services through urban areas.
Steven Hancock, Karen Anderson and Kevin J. Gaston