Can Desert Locust be controlled?
Can we further improve Desert Locust's preventive strategy?
(conference given during the 12th international congress of the Orthopterology, Ilheus, Bahia, Brazil, 31 October - 3 November 2016).
Of course, everybody knows that the Desert locust is an important pest in Africa and Asia. That the invasions may have devastating consequences on food security. That the social consequences can be visible on the long term, even 20 years after an invasion, as it was recently demonstrated. However, over the years, a lot of progress has been made to tackle this insect and to try to prevent the invasions. We now arrive to a more acceptable situation than 50 years ago, and even than 10 years ago. But, is it possible to do better ?
First, I have to say that the system to manage this insect is now considered as a complex one. A system with many components, from field teams to international donors, from small rural communities to pesticides companies. All are expected to try to resolve the locust problem, but the interests are diverse and sometimes contradictory.
What is the current strategy to manage this insect ? How to prevent this solitary Desert Locust to gregarize, to turn into a horde of devastating and gregarious individuals? The current strategy to manage the Desert Locust infestations is continuous survey in the outbreak areas, early warning and preventive control. Which is not so simple when the area to cover is of million of square kilometers.
Anyway, the preventive strategy prove to be efficient. This strategy reduced the magnitude of Desert Locust invasions during the last 50 years, as we can notice on this map which is a comparison of the locust situation during the period 1930-1964 and the period 1965-1999. But this strategy also reduce the frequency and the length of these invasions. Is the current situation a perfect one ? Surely not. Some upsurges and invasions are not stopped at an early stage. So, the question is « Can we further improve this preventive strategy? »
There are several possibilities. They concern the locust, à its environment à and the various stakeholders, the various agents involved in the management process. I shall just take some examples from recent researches, mainly developed by CIRAD in collaboration with FAO and the anti-locust centers in the various affected countries.
For the ecology of the desert locust, several recent studies provided interesting results about the dynamics of the solitary populations, their migrations and the thresholds for gregarization. First of all, a recent study coordinated by Marie Pierre Chapuis. The desert locust is well known in its gregarious phase. But we have a lack of knowledge on the solitary populations, migrating at night, and often unnoticed in the immensity of the Sahara. For several years, we organized a huge operation to collect solitary populations during a remission period, in collaboration with all the locust centres and with the support of the FAO. With all these samples, it was possible to study the genetic structure of the populations. The results were both surprising and novel. Contrary to what might have been expected, almost all the solitary locust populations showed a high genetic diversity. And they did not show any genetic structure, from Mauritania to Pakistan. I will not enter in the details of this study. But the main conclusions were:
that solitary populations are much larger during remission periods than what was previously thought;
that solitary populations have a high mobility ; they certainly migrate intensively, and are thus regularly mixed across the whole of their remission area.
This is a new vision of the solitary populations: larger and migrating intensively.
New interesting results have also been obtained by the analysis of archive data. A lot of historical data is available on the Desert locust. I will comment on a recent study coordinated by Mohamed Lazar in Algeria. Data collected during 30 years were used to analyze solitarious population dynamics across the Sahara where some outbreak areas are located. Each small black dot on this map is a survey point. We analyse more than 30.000 records over a 30 years period. Among many results, it was possible:
to clearly demonstrate the importance of runoff to create suitable habitats over a long period and to very distant places from rain areas.
to establish a statistical relationships between various locations, demonstrating a clear regional dynamics, and confirming the importance of migrations of solitarious populations within the recession area of this species.
The operational implications are multiple. The most important result is probably that this study clearly illustrates the importance to target solitary locusts for a more effective monitoring and more effective preventive actions.
The density threshold for gregarization is another important point to improve the preventive control. This threshold was not yet established. Important results have been obtained recently by Sorry Cissé. Field sampling was carried out in breeding areas of the Desert Locust in Mauritania during two successive years. Hopper densities were assessed at numerous sampling sites. Vegetation was also sampled to characterize the habitats. The results provided a critical density value around 2.45 hoppers / m2. For the adults the density threshold for phase transformation is around 680 adults / hectare. Above this density, gregarious hoppers are expected to be seen more frequently in nature. Of course vegetation cover and height are also important. But hopper density was confirmed as the main factor explaining the presence of gregarious individuals. All these results are important and should help the management of locusts and to the decision making during control operations.
The second line of research is about the environment of the Desert Locust. An ongoing challenge is to be able to guide where prospection surveys should occur depending on local meteorological and vegetation conditions. But the areas of potential gregarization for Desert locust are large. Remote sensing has long been used to try to detect in real time the appearance of ecological conditions for desert locust breeding. Until now remote sensing was used mainly to try to detect rains and areas of green vegetation. Recent satellites should even allow access to a key factor: soil moisture. But can we do better?
A recent study coordinated by Cyril Piou shows that most certainly the answer is yes. This work combines remote sensing data and, again, the archive data accumulated on the Desert Locust over many years. On this map each point is a field survey . The study analyses the relationship between historical survey data and a vegetation index gathered by remote-sensing. It was possible to transform that information into a predictive model of probability of finding locusts on the ground in specific areas. The model automatically transform new NDVI maps into predictive maps of presence/absence of Desert locust. This methodology will certainly help to focus more efficiently the survey efforts on specific parts of the outbreak areas based on the predicted probability of locusts being present. This will certainly greatly reduce the costs of locust survey operations.
Third line of research: the behavior of the actors of locust control. It is also an important component to be considered and to try to better understand. The recent invasions have shown that one of the key problems is the degree of organization of the various agents of locust control. Their ability to quickly mobilize funds, their preparedness to emergency management, etc. As I said above, desert locust management is a complex system. How to optimize the operation of such a complex system ? That is where the multi-agents modeling may play a role.
The purpose of such a multi-agent model is to simulate the actions and interactions of various autonomous agents to assess their effects on the whole system. Such models are increasingly used in plant protection. Such a model was recently developed by some CIRAD colleagues to better understand and to better take into account the behavior of the various agents of locust control. And therefore to assess whether their actions were optimal for an efficient locust control. Such a model take into account:
the natural stochastic events of the development of a locust invasion
the various levels in the management system including a funding institution, a national control unit and survey and control teams.
In fact, initially, the model is not specific to the Desert locust. This is a generic model mainly developed to demonstrate the feasibility and benefits of this approach. Anyway, promising results already have-been obtained. For instance, it was possible to demonstrate, in a parsimonious way, that a dynamic of progressive failure to remember the importance of preventive management is enough for generating the invasion cyclicity observed. It was also possible to demonstrate that the funding institutions could clearly make the preventive management more efficient by increasing by just few percent their base of support to the control units. Only 10% of increase of the allocated budget may reduce the invasion by one-third. We know that more advocacy is needed to support a faster involvement of the donors in Desert locust management. Such a model could be used to convince donors - national or international - to provide a greater and faster support in case of a crisis but also on the long term.
In conclusion, yes we can further improve the efficacy of the preventive control strategy. This is the goal of the ongoing research. And surely, soon, the Desert locusts will be more and more in trouble.
