Monitoring of Vegetation Condition

Example of drought detection by combining precipitation, vegetation indices and satellite derived evapotranspiration


Drought is a recurring climate feature that results from water deficit (precipitation, surface water or groundwater) over an extended period. Meteorological drought usually precedes other types of droughts and occurs when there is a prolonged time with less than average precipitation. Agricultural drought develops when meteorological drought leads to a soil moisture deficit that limits water availability for crops. Hydrological drought and its effects take longer to appear and are a consequence of significantly lower water levels in rivers, reservoirs and aquifers because of long-term precipitation shortages.


Drought indices or indicators represent different components of hydrological cycle (e.g., precipitation, soil moisture and surface or sub-surface water levels) and are very useful in locating drought-impacted regions. In the scope of drought monitoring, some of the well-known indices are Standardised Precipitation Index (SPI) and Palmer Drought Severity Index (PDSI) as indicators of meteorological drought, soil moisture and evapotranspiration indices as well as various indicators of vegetation condition based on satellite data (e.g., VHI, NDVI and FVC). For easier interpretations of drought impacts, it is particularly helpful to analyse statistical anomalies of the current situation with respect to the long-term climatology.


Figure 1 - FVC, actual and reference evapotranspiration are displayed for a single satellite pixel in SE Slovenia in 2017. Additionally, different drought levels are indicated in the figure.


As was demonstrated in Satellite Derived Evapotranspiration Is Advantageous for Drought Detection, a combination of satellite evapotranspiration and vegetation signal is advantageous for drought detection. The advantage of combining evapotranspiration and vegetation for drought monitoring is twofold. Firstly, the ratio between actual and reference evapotranspiration (ET RATIO) gives an earlier warning for the developing drought as drought signal from evapotranspiration usually precedes that of vegetation. Secondly, when both vegetation and evapotranspiration point in the direction of drought conditions, the drought signal is more reliable as it is coming from both sources. Figure 1 illustrates a case when signals from both evapotranspiration and vegetation indicated drought. A moderate drought occurred in the summer months over agricultural areas in SE and E Slovenia in 2017. A single satellite pixel is shown for a representative location in SE Slovenia. The land type of the concerned location is mostly agricultural land with some pockets of mixed forest. Fraction of vegetation cover (FVC, LSA-421), actual evapotranspiration (DMET, LSA-312) and reference evapotranspiration (METREF, LSA-303) are displayed in the figure. All data are obtained from the MSG satellite and are processed within EUMETSAT’s LSA SAF.


ET RATIO < Threshold ET RATIO   ET RATIO < Threshold ET RATIO   ET RATIO < Threshold ET RATIO  
or or or
SPI < Threshold SPI SPI < Threshold SPI SPI < Threshold SPI


and and
  FVC anomaly < Threshold FVC FVC anomaly < Threshold FVC 
  (MSG or METOP) (MSG and METOP)
    FVC anomaly < 1.5 * Threshold FVC
    (MSG or METOP)



Drought levels

Individual indices could be combined in a way that they highlight different drought levels. A useful drought index can be constructed by merging precipitation anomalies (SPI), ET RATIO and FVC anomalies from MSG and Metop satellites. SPI is an indicator of meteorological drought while ET RATIO and FVC anomalies are more associated with agricultural drought. Methodology applied is similar to the one used for JRC’s European Drought Observatory (see The assessment of drought severity is based on the analysis of individual drought indices. There are three drought levels, which are defined whenever indices exceed individually defined thresholds and correspond to different stages and severity of droughts. Level 1 is associated with rainfall and soil moisture deficits. Level 2 corresponds to developing drought while level 3 means a fully developed drought. The basic concept is illustrated on the box on the left.




By combining precipitation, vegetation and evapotranspiration, a more robust and reliable evaluation of drought impacts can be made. Here is an illustration of possible combinations for drought evaluation over Slovenia. SPI is a reliable indicator of meteorological drought but its response to the drought is normally quite slow (SPI calculated over 3-monthly window often works out best for drought monitoring). Evapotranspiration based indices are good at offering an early warning for agricultural droughts. See Evaporative Demand Drought Index (EDDI) as an example, The advantage of vegetation indices is that they provide actual vegetation conditions on the ground. However, they usually indicate drought conditions later than evapotranspiration based indices and they might show vegetation damage that is caused by other meteorological phenomena than drought.