Original Note Outline

• Photoshop a newspaper headline that says "Engineering Solution Saves World" with a picture of an antenna.
• Describe some of the challenges in ecology
  • Resources are limited. It's expensive and difficult to send skilled researchers into the field to gather high quality data. Methods must remain consistent.
  • Wild animals pose a great challenge - they're often difficult to find, difficult to recapture, and move around a lot. Collecting good data can take many years. Cite Driscoll Eyre Peninsula study, perhaps - number of volunteers * trips * years
  • Many ecological events happen rarely, especially dispersal and migration. Unless recapture rates are high, it is virtually impossible to detect them over sensible time scales.
  • Statistics are more reliable with higher samples. There is a big trade-off to be made between the number of samples per individual and number of individuals sampled (some charts with data points). If there aren't enough individuals, how general are the results? If there aren't enough samples per individual, do we really know what it's doing between samples? Does it matter if the samples aren't taken at the same time for every individual?
• Holy grail - automatic sampling
  • Telemetry has allowed us to increase our knowledge of larger animals on global scales. Can track exact position with GPS, or infer it with the period of sunlight.
  • These technologies are impractical for smaller animals that have small home ranges. GPS isn't yet small enough.
  • Radio tracking is widely used but has many problems. Maybe find some studies and show how many individuals were tracked - eg Murray and Shaw was about 9.
  • Automated telemetry - mention C Barro island, Panama. That other bird one with presence absence.
  • Social networks eg ZebraNet, EncounterNet, Markham - still currently quite big, quite expensive, give information on encounters between animals but not always absolute position. This may be sufficient for certain ecological questions.
• Direction finding technology
  • Traditional method - directional antenna, receiver, walking in the rain.
  • Rotating antenna systems - Sirtrack - $85000 for three stations, giving what sort of total area coverage?
  • Multiple fixed antennas, eg C Barro. Good directional antennas are expensive.
  • Harmonic radar. Very cheap transponders, but most ecological studies have used expensive ($10000) off-the-shelf radar equipment that is designed for avalanche rescue and the range is limited to about 10 metres at best. This can be improved upon with a custom design - talk about the insect one. Very much larger installation, but probably quite expensive, and not being able to differentiate individuals may be a problem when answering some questions.
  • Doppler Direction Finding - the technology is relatively cheap, doesn't require expensive antennas or moving parts, and can give precise bearings to transmitters, though accuracy is still an issue. The range and sensitivity is worse than using specialised direction finding equipment such as Yagi antennas, but it's half the price in volume.
    • How does Doppler DF work, briefly?
    • Can also fall back to presence/absence of animals within the vicinity of the base station using the same equipment (with no directional information, though received signal strength could be used as a very rough proxy)
    • Base stations form a sensor network and communicate with each other to triangulate animals and alert each other about the presence and movements in order to take simultaneous fixes, increase and decrease rate of scanning of different channels to maximise data capture.
• Plan B - if cost/benefit isn't worth it, sites can be fitted with only presence/absence detectors, radios are cheaper.
• Ecological question
  • What are some of the interesting questions?
  • Dispersal in fragmented landscapes - could talk about the problems with data in my honours project.
  • Can we see any evidence for metapopulations in a real system?
  • Some of the other things Don mentioned
• Timeline
  • Limited prototype - modelling the behaviour of the system, writing software
  • Limited test in field conditions, perhaps with fake animals, 3-5 units.
  • Full test in a large scale deployment - 10 * 10 KM.

Chapter by chapter / paper summary