Reconstructing a continuous record of internal water content from sparse destructive vegetation sampling and hydrometeorological data

Vegetation water content (VWC) refers to the total mass of water stored inside the above-ground components of the plants per unit land area [1]. VWC is a function of both plant biomass and plant hydraulics, as transpiration and root water uptake govern the withdrawal and recharge of stored water. Consequently, VWC is interesting in a range of applications, from tracking and studying plant development, phenology and potential fire fuels, to the onset and propagation of droughts and water stress.

In recent years, passive and active microwave remote sensing have been employed to estimate VWC, e.g. using vegetation optical depth (VOD), due to the sensitivity of microwave observables to plant water content. The validation of these estimates requires reliable ground data, commonly obtained through destructive vegetation sampling. However, destructive sampling is labor and time-intensive, in particular when sub-daily to seasonal variations are studied.

Here, we present a methodology adopted from the tree physiology community to reconstruct a time series of continuous VWC, using 15-min data of sap flow, root water uptake, transpiration, and sparse destructive sampling. Results from two intensive field campaigns in corn fields are used to demonstrate the potential of this method to capture the internal water dynamics of a crop canopy. Finally, to illustrate the value of reconstructing such datasets, the reconstructed VWC series will be compared to high temporal resolution backscatter data from a tower-based scatterometer and used to interpret sub-daily variations in observed backscatter.