Species Spread and Producer Responses

1. Species Spread and Producer Responses David Adamson and Daniel Quiggin

2. Background • Pest management is part of the decision making process for producers • Crop choices and pest control decisions made jointly • $2,342 million spent by landowners annually on pest control (ABS, 2008)

3. Background • Pests occur in an agricultural landscape • Hence the need to understand how they move and what impact they can have • Until both the distribution and density of a pest through time and space are understood, management expenditure may be misallocated (Adamson, 1996)

4. Literature • Chalak, Pannell and Polyakov (2011) • Optimal level of pest control • Epanchin-Niell and Wilen (2010) • Optimal method of pest control

5. Limitations of existing models • “Rook Contiguity” – pest can move in four directions (North, South, East, West) • Pests can only spread to adjacent cells • Uniform density of infestation

6. Our Model • Removes rook contiguity • Pest can move in nine directions, including worsening the infestation in the origin cell • Non-uniform density • Pests have four levels of infestation (None, Low, Medium, Heavy) • Maximum range of pest increased to five cells • Two commodities • Only one subject to pest damage • Control taken as given • focus is on the effect of implementing the ability to switch commodities

7. Pest Spread • Pest originates in a single cell (with low density) • Spreads to surrounding cells with an assigned probability (can also worsen its own infestation with the same probability) • Once new cells are infested, they also have probabilities of spreading • Eventually, all cells in landscape subject to high-density infestation • This is the equilibrium state with control expenditure • Model yields estimated cost of pest introduction

8. Key concept: Jump-Point Diffusion • With a low probability of spread, it can happen that pests end up far from the origin with nothing in between • Blackberries via birds • In this way, most of the landscape can be covered with low-density infestations • Once a cell becomes heavily infested, the spread can be rapid • Bypasses border controls

9. Economic Evaluation • Landscape has two commodities • Could also represent different land conditions • Each cell represents one unit of one commodity • Cells have different values depending on infestation • Variable of interest is value of landscape, measured by aggregating value of cells • Gross Margin approach

10. Control and Switching • Pest introduced • Only affects one commodity (say, potato bug in a landscape of potatoes and tomatoes) • Paper compares two scenarios • Control only (producers incur fixed control costs) • Control and switching

11. Parameters • Cells valued as in table • Decline in value of Commodity 1 reflects cost of pest control • Producers will switch when the benefits of controlling the pest are outweighed by the (opportunity) costs • Economic Threshold (Headley)

12. Results: Mean Simulated Land Value

13. Results: Standard Deviation of Land Value

14. Discussion • Analysis without commodity switching shows sharper decline and greater variance • In reality, farmers do switch commodities • Cost-benefit analysis without this option may thus be misleading

15. Future Directions • Incorporating cost of capital • Net Margin vs Gross Margin • Monitoring and eradication • Interaction with Jump-Point Diffusion • Stochastic control • “Pests With Benefits” • Hunting or harvesting vertebrate pests

16. Species Spread and Producer Responses David Adamson and Daniel Quiggin