Can satellite data help fight locust outbreaks?

Por GRID-Arendal

22 de July de 2020

East Africa recently experienced its most serious outbreak of locusts in decades, threatening the food security and livelihoods of millions of people. Could Earth observation (EO) satellite data be used to track the locusts? The Fragility, Conflict and Security (FCS) consortium – part of the European Space Agency’s Earth Observation for Sustainable Development (EO4SD) program – recently conducted a rapid assessment to help answer this question.

Exceptional and prolonged rainfall covered most of East Africa from October to December 2019. The unusually wet conditions, along with abundant vegetation in these areas, favoured the development and spread of desert locusts. By early 2020, swarms of locusts had spread into several countries, including Somalia, Kenya, Ethiopia, Uganda, and Tanzania, widely infesting cropland and pastures.

Locust upsurge from Summer 2018 to Winter 2019/20. Source: FAO Locust Watch.

A one-square-kilometre swarm can consume the same amount of food as 35,000 people in one day, and swarms can cover several hundred square kilometres.

Technologies and tools that can better monitor locust movements and impacts are needed to provide leaders with information and to support activities to limit the damage to communities.

Satellite EO technology provides comprehensive information on vegetation vigor and stress, climate, soil moisture, and other variables. Could EO data and spatial analysis provide information on locust presence by monitoring impacts to vegetation?

The FCS consortium completed a rapid evaluation to assess the use of EO data to detect the direct damage of the locust infestation on croplands in Ethiopia. The assessment used data from the UN Food and Agriculture Organization’s (FAO) eLocust3 digital survey system, which covers all locust-affected countries in Africa, the Middle East, and Asia. The system collects field observations on locust presence, abundance, and life-cycle stage and on soil moisture, vegetation status, habitat type, and treatment control.

Data collected in East Africa confirmed a rapid expansion of locust swarms across Ethiopia from August to December 2019. Many upsurges of locusts were recorded in November 2019 near the city of Jijiga, in the eastern Somali region of Ethiopia. The figure below illustrates the spread of the swarm through Somalia into eastern Ethiopia between August and December 2019.

Can satellite data help fight locust outbreaks?

The area north of Jijiga is a relatively flat landscape with rolling hills characterized by a mosaic of croplands and sparse vegetation. The upsurge of swarms in this region originated from the north and east. By the end of November 2019, more than a dozen swarms had been observed through the eLocust system, as shown in the image below.

Cropland north of Jijiga, Ethiopia. Source: Google Maps, November 2018.

Sparse vegetation north of Jijiga, Ethiopia. Source: Google Maps, November 2018.

The FCS consortium analysed a series of EO images of vegetation vigor taken at different times to see if it could detect impacts from locust swarms. The images showed that there had been changes, as illustrated in the figure below, but it was not possible to determine if these changes were due to seasonal vegetation differences or the locust swarm invasions.

To try to determine if the findings represented a continuation or deviation from historical trends, analysts calculated median values from the same time period from 2015 to 2018 and compared them to the 2019 findings. The results are illustrated in the figure below.

The analysis showed small reductions in vegetation vigor between 2019 and previous years in many of the areas where locust swarms had been observed.

However, the decreases cannot solely be attributed to locusts because seasonal decreases in vegetation vigor normally occur in November and December, and normal harvesting practices decrease vegetation vigor too. The area also had abnormally abundant vegetation in 2019 prior to the locust infestation (see the graph below), making it more difficult to separate locust impact from normal changes in vegetation.

Still, the FCS analysis illustrates the ability of EO data to identify vegetation anomalies through time series analysis. Identifying areas of abnormally high vegetation compared to historical averages could provide strategic information to support locust-control activities.

Further investigations of vegetation anomalies from 2014 to 2019 using the same method with other sources of satellite data found that swarms may be moving to areas with unusually high vegetation growth, during a period when vegetation is generally senescing across the region. The image below highlights these anomalies; areas where vegetation growth exceeds the historical average by 20 per cent are outlined in dark green.

The FCS analysts concluded that while it is challenging to detect and identify direct locust impacts, time series analysis of EO data may be useful in identifying areas with unusually high vegetation growth that are at higher risk of locust infestation. Ultimately, they determined that EO data and derived products coupled with climate and soil moisture data can help in the monitoring of potential locust breeding and swarming areas. These tools could support decision makers in improving response and mitigation of locust outbreaks.

-----

A longer version of this story with more technical detail was originally published on the FCS website. The analysis was completed by the FCS consortium as part of the EO4SD FCS project. The analysts thank the FAO for sharing eLocust3 data and the World Bank staff for providing guidance and feedback on the preliminary analyses. GRID-Arendal is one of the partners in the FCS consortium.