In Situ Calibration of Light Sensors for Long-Term Monitoring of Vegetation

1  Abstract—Light sensors are increasingly used to monitor vegetation growing status by measuring reflectance or transmittance in multispectral or photosynthetically active radiation (PAR) bands. The measurements are then used to estimate vegetation indices or the fraction of absorbed PAR (FPAR) in a continuous and long-term manner, and to serve as inputs to environmental monitoring and calibration / validation data for satellite remote sensing. However, light-sensor calibration is often overlooked or not properly attended to, which leads to difficulties when comparing measurement results across sites and through time. In this paper we investigate a practical and accurate user-level in situ calibration method in daylight. Calibration of a sensor pair is made for measuring either bihemispherical reflectance (BHR) or hemispherical-conical reflectance (HCR), the two most common ground-based spectral measurements. Procedures and considerations are suggested for user calibration. We also provide a method for calibrating and measuring single-sensor reflectance-derived Normalized Difference Vegetation Index (NDVI) from red and near-infrared (NIR) bands. The calibration error propagation is analysed and the induced uncertainties in vegetation reflectance and NDVI are evaluated. The analysis and field measurements show that NDVI estimated from a user calibration factor can be as accurate as, or even more accurate than, manufacturer calibration. The in situ calibration described here remedies the situation that reflectance for large field-of-view sensors cannot always be estimated from manufacturer's calibration. The method developed in the paper may help improving the reliability of long-term field spectral measurements and contribute to the near-surface remote sensing of vegetation.