Inferring Grassland Drought Stress with Unsupervised Learning from Airborne Hyperspectral VNIR Imagery

Abstract

The 2018–2019 Central European drought had a grave impact on natural and managed ecosystems, affecting their health and productivity. We examined patterns in hyperspectral VNIR imagery using an unsupervised learning approach to improve ecosystem monitoring and the understanding of grassland drought responses. The main objectives of this study were (1) to evaluate the application of simplex volume maximisation (SiVM), an unsupervised learning method, for the detection of grassland drought stress in high-dimensional remote sensing data at the ecosystem scale and (2) to analyse the contributions of different spectral plant and soil traits to the computed stress signal. The drought status of the research site was assessed with a non-parametric standardised precipitation–evapotranspiration index (SPEI) and soil moisture measurements. We used airborne HySpex VNIR-1800 data from spring 2018 and 2019 to compare vegetation condition at the onset of the drought with the state after one year. SiVM, an interpretable matrix factorisation technique, was used to derive typical extreme spectra (archetypes) from the hyperspectral data. The classification of archetypes allowed for the inference of qualitative drought stress levels. The results were evaluated using a set of geophysical measurements and vegetation indices as proxy variables for drought-inhibited vegetation growth. The successful application of SiVM for grassland stress detection at the ecosystem canopy scale was verified in a correlation analysis. The predictor importance was assessed with boosted beta regression. In the resulting interannual stress model, carotenoid-related variables had among the highest coefficient values. The significance of the photochemical reflectance index that uses 512 nm as reference wavelength (PRI512) demonstrates the value of combining imaging spectrometry and unsupervised learning for the monitoring of vegetation stress. It also shows the potential of archetypical reflectance spectra to be used for the remote estimation of photosynthetic efficiency. More conclusive results could be achieved by using vegetation measurements instead of proxy variables for evaluation. It must also be investigated how the method can be generalised across ecosystems.

Review of Remote Sensing Applications in Grassland Monitoring

“…The validation of this index was successfully performed at regional scales as well as at larger scales. In 2021, Hermanns et al [186] made full use of hyperspectral images. They combined a standardized precipitation evapotranspiration index (SPEI), a fused index considering potential evapotranspiration and precipitation, with ground-measured soil moisture to represent the drought stress.…”

“…In addition, all these common remote sensing images belong to optical images, which will inevitably be affected by cloud cover and shadowing, resulting in the lack of spatial information and breaks in temporal continuity. Although there were a few studies utilizing UAV or airplane images [23,186] and radar data [19,117,185], it still has failed to fundamentally solve the problem. UAV images have significant limitations on the shooting locations and coverage size.…”

Estimating soil moisture content under grassland with hyperspectral data using radiative transfer modelling and machine learning

“…During the date of acquisition, the entire area was used for cattle grazing. We consider the hydrological status at the site as water limited, since Hermanns et al (2021) detected drought stress of the vegetation at the site at even higher SMC than those measured in our data acqusition ( Hermanns et al, 2021 ).…”

“…Grosses Bruch is a pasture within a flat wetland area ( Wollschläger et al, 2016 ). It is located in the Central German Lowland and susceptible to drought due to negative climatic water balance ( Hermanns et al, 2021 ). The soil texture is defined as highly clayey silt ( BGR, 2007 ).…”

Characterising turf grass adaptation to drought and heat and assessing its potential cooling effect : Combined report on turf grass field experiments : Work package 1: Climate and drought, Work package 2: Climate and temperature

“…For our study, we used a combination of hyperspectral imaging techniques, soil sensor measurements, and gene expression analysis via RNA sequencing, to decipher the research questions. Hyperspectral imaging is an established method that is used to estimate plant biomass, structure, photosynthetic activity and biochemical properties (Hermanns et al 2021;Théau et al 2021;Zhang et al 2021), and therefore has potential use in our studies. Here, we defined drought resilience in the genotypes as monocultures and mixtures that maintain optimal growth during periods of water depravation and additionally, recover quickly from drought (and heat) stress after rainfall or irrigation (and lowering of temperature) occur.…”