Integrating Visual (Ground-Based) and Near-Surface Remote Sensing Phenology in a Deciduous Urban Woodlot

Abstract:

Phenological measures offer important opportunities to improve understanding of ecosystem processes. However, it is typically quite difficult to combine detailed visual ground observations, which provide necessary information on species timing differences, and satellite-derived remote sensing data, which facilitate needed spatial integration and large area coverage. A new resource to address this scaling issue is near-surface remote sensing data collected from fixed position cameras. This paper presents initial findings from a two-year comparison of the spring and autumn seasonal transitions in Downer Woods, a small urban woodlot on the University of Wisconsin-Milwaukee campus dominated by white ash and basswood trees. The study area is under observation from a visible/near-infrared camera installed in March 2013 that is part of the Phenocam network, and also has detailed ground-based species-specific visual phenological observations collected in both spring and autumn, as well as air/soil temperatures and light sensor data measured under the canopy. The results show that at this location, the Phenocam visible/near-infrared band data series can be successfully compared to aggregated species visual phenological observations. Further, both of these changes can be in turn simulated by process models based on seasonal temperature changes. Thus, the concurrent collection of these data suggest a coherent process whereby more robust ground-based species-aggregated “pixel” data can be produced which will be scalable to large areas, and potentially be applicable to more complex environments and ecosystems. Such an approach could potentially improve phenology-based spatial estimates of carbon and energy flux.