Spectral vegetation indices and uncertainty: insights from a user's perspective

The primary objectives of this response communication are to provide insight into the use of spectral vegetation indices (SVIs) and a user's perspective on the uncertainty in SVI values, especially when these are derived from multiple sensors. We review how two papers in this special issue address uncertainty, and we explore two practical applications of SVI products and how comprehensive quantification and spatially explicit visualization of uncertainty could enhance their use. Although researchers identify the causes of uncertainties in SVIs, there has been little advancement in connecting and integrating the associated uncertainties inherent to all steps of the processing and model chains (e.g., data capture, data input and SVI generation). Cross-comparison of uncertainty assessment is challenging to the end-product user because reporting of uncertainty tends to be research or data product-specific with limited emphasis on facilitating the interpretation of uncertainty associated with algorithm and processing quality for use by managers or decision makers. Consequently, the confidence in these data is often based on experience and visual confirmation of the spatial and temporal consistency in SVI imagery and time-series data. Although the level of accuracy required varies depending on use, overall product quality assurance and a comprehensive, site-specific uncertainty assessment bundled with SVI data fields could mean the difference between using SVIs to report on spatial-temporal patterns versus using these data to make natural resource management decisions.