Traditional drought monitoring relies heavily on ground-based weather stations tracking precipitation and temperature. While valuable, these point measurements fail to capture the continuous spatial variations of drought across diverse landscapes. To bridge this gap, modern remote sensing leverages a fundamental physical reality: when the land surface runs out of water, its energy balance shifts. Instead of using incoming solar energy to evaporate water (latent heat), the surface heats up (sensible heat). By observing this thermal response alongside vegetation health from space, researchers can directly monitor environmental water stress.
The LST-VI Feature Space: The Mechanics of the Triangle
One of the most intuitive ways to extract drought information from satellite data is by plotting Land Surface Temperature (LST) against a Vegetation Index (VI), such as the Normalized Difference Vegetation Index (NDVI) or Enhanced Vegetation Index (EVI). When thousands of pixels from a regional satellite image are plotted together, they naturally form a triangular or trapezoidal shape.
To understand why this happens, consider the two extremes of land cover: bare soil and dense vegetation.
- Bare Soil (Low VI): When bare soil is completely dry, it cannot dissipate heat through evaporation, causing its temperature to soar. When it is wet, evaporation keeps it cool. This creates a wide temperature variance at the low end of the vegetation axis.
- Dense Vegetation (High VI): Healthy, well-watered plants transpire water to regulate their temperature, staying cool even under intense sunlight. However, if a drought restricts their water supply, plants close their stomata to conserve moisture. Transpiration drops, and the canopy temperature rises.
The upper boundary of this scatter plot—where temperatures are highest for any given vegetation level—is called the dry edge, representing severe water limitation. The lower boundary is the wet edge, representing maximum water availability.
By calculating where a specific pixel falls relative to these boundaries, researchers derive the Temperature Vegetation Drought Index (TVDI). As shown by Li Wu et al. and Wanqiang Han et al., TVDI provides full spatial coverage that overcomes the limitations of sparse weather stations. Guangzhi Rong et al. demonstrate that this fine spatiotemporal resolution is uniquely suited for capturing agricultural water stress, while Gang Wang et al. even used a calibrated TVDI framework to map localized soil moisture depletion caused by wind farms.
Beyond simple empirical triangles, this thermal-vegetation coupling underpins advanced physical frameworks. For instance, Two-Source Energy Balance (TSEB) models utilize this trapezoid space to isolate the distinct thermal contributions of soil evaporation and plant transpiration, enhancing our ability to detect vegetation water stress (Yongmin Yang et al.). Similarly, the Crop Water Stress Index (CWSI) tracks environmental stress by comparing actual canopy temperatures directly to a theoretical non-stressed baseline (Ruixue Ma et al.).
Evaporative Stress: Tracking the Supply-Demand Gap
While the LST-VI space observes the thermal consequence of water limitation, another powerful approach measures the flux of water directly by comparing Actual Evapotranspiration (AET) to Potential Evapotranspiration (PET).
Think of PET as the atmosphere's "demand" for water—determined by solar radiation, temperature, wind, and humidity. AET is the "supply" that the land surface actually delivers. In a well-watered ecosystem, AET closely matches PET. When soil moisture drops, the land surface cannot meet the atmospheric demand, and AET plummets below PET.
The ratio of these two metrics, often termed the Evaporative Stress Ratio (ESR), serves as a direct proxy for soil water availability (Yongxian Su et al.; Yang Zhou et al.). By normalizing this relationship over time, researchers calculate the Evaporative Stress Index (ESI). As Linjing Qiu et al. and Jizhou Bai et al. demonstrate, because ESI integrates both atmospheric demand and surface moisture constraints, it reacts more dynamically to evolving drought conditions than traditional precipitation-based indices.
Early Warning and Flash Drought Identification
The true power of combining thermal and evaporative indices lies in early detection, particularly for "flash droughts"—rapidly intensifying events driven by extreme heat and atmospheric dryness.
Because traditional vegetation indices like NDVI reflect structural changes (such as leaves browning or dying), they act as lagging indicators of drought. In contrast, evaporative anomalies occur the moment water constraints begin. Li Jia et al. leveraged this by using the Evaporative Stress Anomaly Index (ESAI) alongside cross-spectral analysis to detect vegetation stress significantly earlier than traditional methods. To track these rapid lifecycles from onset to recovery, Li Jia et al. also introduced the Evaporative Stress Percentile (ESP), providing a robust framework across diverse ecosystems.
By integrating these evaporative fractions with surface soil moisture measurements, researchers can construct highly sensitive indicators like the soil moisture stress index (SMSI) to quantify canopy resistance (Chaolei Zheng et al.). Furthermore, refining the underlying vegetation metrics themselves—such as using the Universal Pattern Decomposition Method (VIUPD) to enhance the traditional Vegetation Condition Index (VCI)—ensures that spatial drought impacts are captured with maximum fidelity across complex landscapes (Wenzhe Jiao et al.).
Key Foundational Work
Intercomparison of Three Two-Source Energy Balance Models for Partitioning Evaporation and Transpiration in Semiarid Climates (2018) — contributed 1 concept elements to this direction.
Evaluating an Enhanced Vegetation Condition Index (VCI) Based on VIUPD for Drought Monitoring in the Continental United States (2016) — contributed 1 concept elements to this direction.