Unread postby Lu Rome » Sat Dec 16, 2017 12:40 pm
Historically, it was popular to use fecal counts based either up on the number of pellets or pellet groups found. These were based upon studies that showed how many fecal pellets a deer produced in a day. Of course they soon discovered how variable that was (as you could imagine), how difficult it might be to determine age, etc., and moved on to something else. I'm unsure of how many states actually still use fecal counts. Most have moved on to other techniques.
Spotlight surveys are popular, however they have their own issues. How much land is surveyed, how likely are deer to be visible from a road, are you seeing "visible" deer, can you classify the deer you see and so on.
Aerial surveys have similar issues, but they have the luxury of not being bound to roads and the obvious effects that roads have on animal behavior. Many of the issues can be corrected for by modelling the "visibility" of the animals, and this is done with a preliminary, "calibrating" study to determine the visibility bias. Basically, using collared animals you collect info on the groups of animals that you spot as well as those you miss. With all that info you can then estimate how many animals you missed based upon the information from the groups of animals that you observed.
Most surveys don't have to sample the entire area to calculate an estimate. The models extrapolate from sampling portions. This can still be done even when densities vary greatly across an area using a stratified random sample. Units are broken into smaller portions and areas are sampled and classified into whatever density category they might be estimated to fall into. This can be estimated based upon habitat, personal observations, etc. Much of this is like why states can estimate age structure while only sampling a small portion of the harvest. If you get a large enough sample (which doesn't have to be that large) you can be quite confident of what the whole population looks like, within a certain confidence interval.
FLIR surveys work with similar methodology and can be more accurate, it's just that they are exceptionally expensive. They are much safer than low level flights, but take a lot of time to analyze and can perform poorly under certain circumstances.
Like I mentioned earlier, it isn't always the overall number that matters so much as it is the trend of data from year to year or over a number of years.
I know I'm probably leaving something out and likely botching my explanation of the methodology, but it's been a few years....
I'm surprised QDMA doesn't put out a similar harvest map given they receive a lot of the harvest data from agencies every year.
“Curiosity never killed the cat. The cat died from stupidity, or maybe an overdose of mice.” -The Old Man