Transcript Apresentação do PowerPoint

Applied Computational Modelling to
support studies on agricultural pest
management, considering the
landscape
MSc. Adriano Gomes Garcia
Entomology Department
Agricultural pest management
• Chemicals
• Sterile insect
technique
• Pheromones
• Landscape
management
Effects associated to insecticides
• Insect
resistance
• Toxic effects on
human feeding
• Incompatibility with
local ecossystems
• Rachel
Carlson,1962
Agricultural pest management
• Growing concern on the impact of chemicals
• Integrated pest management: Strategies to reduce insecticide
applications, maximizing natural control.
• It is necessary to understand better the population dynamics on
agricultural systems.
Landscape Ecology
• Is the population dynamics enough to understand all complex
interactions inside an ecosytem?
Challenges in working with the landscape
• High heterogeneity.
• Instability.
• How to represent it ?
Computational approaches
• Allow to create hypothetical landscapes or working with real areas.
• Good approach for studies whose experiment fields are inviable.
• Prediction or identification of patterns: simulations by using
different programming languages (C, C++) or softwares (GIS).
• Example: Cellular Automata.
Cellular automata
• Mathematical discrete model
• Grid of cells
• Cell states change over time steps
• Transition rules
Cellular automata-EXAMPLE
In each time step the adult can lay eggs in 3 neighbour cells
In each time step the adult can die with probability 0.5.
In each time step larvae can emerge from 2 in 3 eggs
In each time step larvae can either die with probability 0.5
or emerge in adult with probability 0.5
How to apply CA in agricultural pest
management?
• Intercropping systems
• Refuge areas
• Control strategies based on manipulation of the landscape
Intercropping systems
• Practice in growing two or more crops in proximity
• Most used arrangement: alternated rows.
• Strategy based on the nutritional ecology of invader insects:
Combining non-suscetible hosts to suscetible hosts.
Intercropping systems
How to simulate insect dynamics in an
intercropping system by using CA
Study case:
• Diabrotica speciosa: Polyphagous beetle
• Hosts: Corn, soybean, bean and potato
• Different fitness for adult and larva stage in each host.
Host
Oviposition(day1)
Larval mortality(day-1)
Larva-adult
development(day-1)
Adult mortality(day-1)
Potato
0.379
0.005
0.027
0.020
Bean
0.394
0.085
0.036
0.020
Corn
0.011
0.011
0.040
0.031
Soybean
0.056
0.045
0.037
0.020
Ávila &Parra, 2002
Oviposition
Larva mortality
Larva-adult development
Adult mortality
Oviposition’
Larva mortality’
Larva-adult development’
Adult mortality’
CA 2: adult dynamics
CA 1: larva dynamics
Transition rules
CA1:
• a) a cell occupied by a larva can become empty with probability μ + α due to
larval mortality or adult emergence , respectively.
• b)an empty cell can become occupied by a larva if an adult lays eggs on it with a
probability β.
CA2:
• a) a cell occupied by an adult female can become empty with probability δ due
to adult mortality.
• b) an empty cell can be occupied with probability α /2 if a larva in the
correspondent cell in CA1 turns into a female adult. The fraction ½ is related to
sex ratio.
Spatio-temporal evolution (larva)
soybean-corn
soybean-potato
soybean-bean
corn-potato
corn-bean
bean-potato
population density
population density
row
row
soybean-bean
population density
population density
soybean-corn
row
soybean-potato
row
corn-bean
populational density
population density
Population density per row (horizontal view)
row
corn-potato
row
bean-potato
time
Average distance
Average distance
Average distance
Average distance reached over the time
time
soybean-bean
time
soybean-potato
Average distance
Average distance
Average distance
soybean-corn
time
corn-bean
time
corn-potato
time
bean-potato
Considerations
• By mean of CA, it was possible to predict the population behavior of
D.speciosa on different combinations of crops in intercropping
systems.
• Corn has shown the better crop to be inserted in an intercropping
system since the population density and dispersion ability were
reduced
Refuge areas and resistance evolution
• Transgenic crop: genetically modified crop
• Cultivation of nontransgenic crops in association with transgenic
crops to manage of insect resistance.
• Computational programming by using celular automata
(methodology similar to intercropping systems
Refuge Areas
Possible study cases
• Helicoverpa armigera and Spodoptera frugiperda: polyphagous
lepidopterous pests that are the main target of Bt-crops.
• Understanding the whole resistance evolution when a new pest
arrives to the agriculture environment would provide importante
results for agriculture
• Incipient project: no results achieved yet
Working with satellite images
• Because of the high diversity in real landscapes, it is
necessary to work with real images (from satellite).
• Geographic information system
Geographic Information System

Hardware, software and data for capturing,
managing, analyzing, and displaying geographically
referenced information.

GPS use.

Georreferenced image.

Softwares: ArcGIS, MapInfo, Fragstat (free).
ArcGIS



Geographic information system for working with maps and geographic
information.
Create, share, and manage geographic data, maps, and analytical models.
Geostatistical Analyst Tools e Spatial Statistics Tools: Regression Analysis,
Krigging,Cellular Automata
Lygus spp : western tarnished plant bug
Local: Cotton field from San Joaquin Valley
Hypothesis: Verify if Lygus hesperus density in cotton fields is
correlated to the density of the same specie in other crops close to
the fields.
• Chilo partellus is one of the main lepdopterous that attack maize
and sorgum
• Cotesia flavipes is a promise for biological control since it is a larval
endoparasitoid of C.partellus.
• Objective: Predict distribution of C.partellus and C.flavipes in all
Ethiopia.
Final Considerations
• It is importante to understand how landscape elements Interact
with insect populations.
• Computational approaches are useful to represent and analyse
landscape factors.
• There is still a great potential to work with computational modelling
in landscape management for controlling pests.
Research group
Profª.Drª.Cláudia Pio Ferreira
Prof.Dr.Fernando Cônsoli
Prof.Dr.Wesley A.C.Godoy
Thank you!
webmail: [email protected]