Print PageWe use spectral reflectance to monitor plant life history. The primary metric is Normalized Difference Vegetation Index (NDVI), a "greenness index". This is based on the ratio of visible to near-infrared. Plants absorb visible light to use for photosynthesis, but have evolved to reflect as much near-infrared as possible because it is of no use to them.
The spatial scale of NDVI monitoring has ranged from continent-wide greenness mapping by satellites to sensors that clip on an individual leaf. Our studies cover a middle ground. We log visible and near-infrared reflectance, over the growing season, from patches of landscape such as meadow or tundra. The patches are on the order of a square meter in size. In these season-long NDVI traces, we look for plant life history inflection points such as peak greenness and onset of senescence. Over a period of years, shifts in the dates these inflection points would indicate climate change.
Results/Conclusions
In early prototypes we have used off-the-shelf weather station components to record the spectral data. These systems have consisted of irradiance sensors connected to commercial general-purpose data loggers. As interest in our methods grows, we are developing versions that are "embedded systems", specialized microprocessor-based electronic circuits tailored to the specific task of irradiance logging. We are realizing order-of-magnitude savings in cost and weight. This will allow scaling up our NDVI monitoring, and make it more feasible for other researchers to collaborate and use our methods.
Season-long irradiance logs are huge data sets. Daily weather causes large fluctuations in sky irradiance, and thus introduces a great deal of noise. We have developed algorithms to extract season-long NDVI traces from the background noise. These traces successfully illustrate the plant life history inflection points.
Data should flow as effortlessly as possible from collection, through collation, to data storage and analysis. In line with this, we are migrating our data tools to the web. We are packaging our analysis, so data sets will be available from online databases using browser-based "data menus". This facilitates collaboration within our group and also makes it easier for other researchers to work with us and use our methods.
