Towards A Community Protocol for Validating Global Climate Data Records Representing Sub-Pixel Land Cover Components and Changes Through Time Using AVHRR, MODIS and VIIRS

As part of the activities of the MODIS Science Team, a workshop was conducted under the auspices of the Working Group on Calibration and Validation (WGCV) of the Committee on Earth Observation Satellites (CEOS) and the Global Observations of Forest Cover and Land Dynamics (GOFC-GOLD) to evolve standards for land characterization and validation of Climate Data Records (CDR), which quantify sub-pixel land cover components. The authors present results from the workshop and a strategy to attain international consensus on validation procedures of sub-pixel land cover data sets derived from polar orbiting earth observatories. The primary focus of the effort is to develop a community consensus protocol for this validation which can be easily applied by researchers worldwide. Precursor work with the MODIS Vegetation Continuous Fields (VCF) product employed a combined remote sensing and field observation methodology which directly ties observed tree canopy closure measurements at ground level to high resolution imagery and in turn to moderate resolution data products derived from the AVHRR and MODIS instruments. The meeting resulted in consensus on the key issues needing to be resolved for validation of sub-pixel vegetation cover estimates both for use as one-time map layers and as multi-temporal change products. Key issues identified were: Concepts and developments in VCF mapping; VCF validation requirements and approaches; Scaling of VCF information from in situ to global measurements; Links between VCF and other approaches in land cover mapping and validation; and Fostering implementation for the NPOESS operational terrestrial observatory. The proposed validation protocol is constrained by, and provides recommendations on: Appropriateness and availability of ancillary data; Scaling via intermediate imagery products versus a direct correspondence with in situ observations; Practicality, repeatability and cost; as well as the appropriate tessellation, aggregation, and collection of long-term ground observations. Initial consideration was also given to experimental validation protocols for areas where the land cover is not woody. Measurements of interest include sub-pixel estimates of crops, herbaceous cover, leaf type, leaf longevity, water cover, and bare ground.