I'm not sure how many of you analyze your trail camera data, but I have a question for those who do and for the camera experts out there. Look at this chart:



It represents several years worth of data and you can see the normal seasonal shifts one would expect. There is not enough information on this chart to do any real year-to-year analysis since the number of cameras change from year to year.

The X axis represents the day of the year and the Y axis represents a 24-hour clock on this scatter plot. The two yellow lines represent the beginning and end of shooting time at this location. Each data point represents a picture with one or more deer in it. Most all of these cameras are on fields.

OK, the thing that caught my interest is the horizontal banding. It is especially noticeable in this year's data during the summer, but I think it is there in the fall and winter as well but there is just not enough data to make it obvious. This banding looks like it is about between bands give or take a little. I can't think of any technical camera related reason for this, but maybe I'm missing something. I think I'm seeing real deer activity periodicity. I've submitted this to a couple of biologist friends for comment, but I thought I'd see what the consensus was from you guys in the mean time?

Is one hour a typical cycle time for deer activity due to digestion or something? Is there some camera related thing that could cause this banding?

Thanks,

Jack

I have never put my pics through a program like yours. Definitely more activity during the growing season. Not sure what would case the banding. If this is from all your cams at different locations, I don't think you have cam issues. Must be some internal clock in the deer and their movement. Very interesting. My 2cents worth.

_________________
Dave,
NRA Life Member
Outdoor-Talk
BlackHillsDave Trail Cam Photos
My Best Buddy,

I agree with Blackhills. Months of ample food source and low pressure increase the prevalence of movment and chronological times of movement and/or browsing. I am no Biologist, but the recurrence (periodicity) of deer movement seems to band in a smaller micro bands (1 hour) give or take the 5 hour band of mass movement. These cams, and I know you mentioned it, must be primarly based in fields. The months of low prressure and high food yields...along with no major pressure at those calendar times seem accurate on this scattergram. Also, is pressure high or low in these areas year round, during hunting,...? I do not think it is cam relative. Geez Jack...awesome stuff, but definitely more than I want to do

P.S. Are these the frig cam's data?

_________________
Its not the kill, its the thrill of the chase

Yes, you guys are right about the seasonal stuff. These cameras are mostly on food plots. I've got other graphs that show the seasonal shifts much more clearly. Basically, deer slowly react to increased pressure during archery season through muzzle-loader. There has been a huge drop during firearms season since I'm in a county that allows deer hunting with dogs. Once they release the dogs, our deer head for heavy cover, lay down, and become largely nocturnal.

We've been working for about 5 years now with prosecutions, fencing, etc to significantly reduce trespass and poaching. As we have done this, it has stiffened the spines of some of our other neighbors who were putting up with the illegal activity. This has pushed the release of hounds further away in general. We still have them but in fewer numbers and less frequently. This year we had a mast crop failure. Our property is a pine plantation with a pipeline running through it. Our food plots were the only quality food around (other than the illegal baiting) with no acorns. So this year we had a huge increase in in-season sightings both by our hunters and cameras. It takes our deer a while after the season's end around the first of the year to become comfortable again. By that time they are on browse and there is not much left in our green fields, so camera events don't really pick up significantly until spring.

It sure looks to me like the banding which is most evident where there is the most data (low pressure, high food value summertime), it deer activity related. If it was on a 3 hour cycle or so, I'd just accept it. I'm not a biologist either and I'm consulting some, but I've never heard of a one hour activity cycle, so I'm suspicious.

I was hoping that either someone had seen this before, or there is something stupid I did that I'm missing.

Thanks,

Jack

Good questions. With that information, it makes a little more sense. The drive through method is what I call it. In my area of Ohio, the deer are plentiful and the season bag limit has gotten liberal, yet so has the hunters and poachers. It seems that you can not hardly find good herd ratios or mangaged areas anymore without a fat wallet. Thus,the hunting pressure increases and the deer become drive thru eaters...you know get it and go, or come back before closing. Quick and meaningful browsing by mass, yet alot of recurrences at limited time and in spurts. Good stuff again Jack.

_________________
Its not the kill, its the thrill of the chase

WOW, Jack, that is some very, VERY cool data!

The banding is obvious, but what is driving that is beyond me. I've seen nothing in the scientific literature concerning this type of periodicity, probably because no one has ever collected data at this level of detail. Even the most intense radio/GPS-collar studies only collect deer location/activity once per hour, which would be far too infrequent to highlight this behavior.

You may be on to something new.

However, the influence of hunting pressure in driving deer off your food plots is pretty obvious too, and a well-understand behavior!

_________________
Bryan Kinkel
BSK Consulting
Specializing in small land management

Very interestin'...

Ya'll look at the chart and see seasonal movement patterns (me too) and "banding" (what's banding I thought that what they did to birds to track there movement ) BUT what I see that most interestin' is in 2009 ya have almost as many deer picture than in the other years listed... That is unless red cover all other colors... If not then what does that lead ya to believe.. 1) deer population "exploded" in 2009... 2) deer got used to the cameras... 3) deer moved from neighbor that was usin' white flash and red blob flash cameras...

Ya'll throw out some ideas... Maybe my cypherin' ain't too good but ain't they a lot more red dots (or whatever angle that is) than yellow, green or blue?????

_________________

The Secretary of Defense of the Chasingame Official Thread Hijackers Union
See ya and God bless;
Phil "Ol' Arky" Weaver

BSK,

Yes, I can think of nothing to account for this. I emailed a couple state biologists. Maybe they can think of something. The one hour periodicity is the thing that I just can't account for. I almost wish the seasonal trends were not there so I could see if the banding conforms to the sunrise/sunset or stays flat. If it conforms to sunrise/sunset, it has got to be a real deer activity phenomenon.

Arky,

Yes, I warned you can't read too much in to year to year changes with this data. First, I have more cameras each year and since I went to BEC, I don't have the down time for sending cameras back. Second, we are on a pine farm. Although we have about 80 acres in mixed hardwood/softwood riparian buffers, the neighboring properties have many more hardwoods. We had a mast crop failure this year. My camera are mostly on food plots and our plots were the only high quality food available for miles except for the illegal baiting. Finally, the charting software does put the red symbols on top.

I also have hunter observation data. It shows a factor of 2 increase in deer observed by hunter hour, so there is a real increase in deer visibility (and probably numbers) this year, but they chart you are looking at my over emphasize that.

Thanks,

Jack

This is probably a stupid question but that never stopped me before. Since these cams are on food plots is it possible that they are feeding for an hour, then chewing their cud for an hour, etc.? I wouldn't think that the times would be so consistant from day to day but if they enter the food plots at the same time maybe it is.
I don't know, just tossing out ideas to chew on.

Great stuff Jack. When using data from trail-cameras you always have to be cognizant of potential data bias, and lots of things can cause bias: the type of cameras used, the number of cameras used per unit area, the settings of each camera (delay and number of repeat photos), and camera placement. In addition, food source and/or hunting pressure changes from year to year will also cause major changes in deer behavior.

_________________
Bryan Kinkel
BSK Consulting
Specializing in small land management

Good one Trapper!

Hey, anything is possible.

_________________
Bryan Kinkel
BSK Consulting
Specializing in small land management

Actually brencade, it's been postulated that that is exactly why ruminant biology first evolved--because it limits exposure to predators in open areas. A ruminant can wolf down food in open areas quickly and then retreat to cover to chew it later.

_________________
Bryan Kinkel
BSK Consulting
Specializing in small land management

Trapper,

I hope that is not a stupid question because it is exactly the one I posed to the biologists. My first thought was that it is related to the digestive cycle. I just got the first biologist reply and he echo's Bryan saying he knows of no biological reason cited in the literature.

Bryan,

You are absolutely right and I've been racking my brain for any kind of bias that could cause this banding. Most of the cameras are BEC Orions running wirelessly, but I do have a few older standalone cameras. The process I use is this:

1) My field cameras are mostly programmed in burst mode with 15 seconds between bursts and set to the minimum delay between triggers. I did put the cameras in video mode during part of the season this year to look at some behavior, but none of that occurred during the summer when the banding is most evident.

2) I extract the pictures and store them on my hard disk. I go through them manually and pick the best pic of a burst series or back to back triggers of the same deer within a couple minute window. For does, fawn, young bucks, and unidentified deer I delete all but this selected picture. For possible shooter bucks, I retain multiple pictures if they help in rack evaluation. (This could cause a bias toward shooter bucks if you used the raw data without care, but nothing related to this chart).

3) I wrote software that extracts the date and time from the EXIF data stored in each picture along with other camera parameters. These are then stored in an Access database along with a link to the picture.

4) I then go through each database record, display the picture, and record in the database record the count of different kinds of animals in the picture. For deer, I record the count of Bucks, Does, Fawns, and Unidentified deer. I also record Shooter Bucks but this number represents a subset of the number in the Bucks field.

5) The chart is produced from querying the database for records with Bucks, Does, Fawns, or Unidentified deer counts greater than zero. The number of deer is not used in the graph, just the date and time of the pic and the fact that it contained one or more deer.


Thanks,

Jack

After reading these posts, and while talking to my biologist friend, I thought one possibility might be some kind of plotting artifact causing this. So, I zoomed in on the data from midnight to 0700:



It was clear the problem was not a charting issue since the bands were very evident. In fact, the dead spots were so clean I knew this could not be real.

Mystery partially solved:

Seeing this, I went back to the database and did a query looking for deer pictures between 0350 and 0400 a completed dead band on the chart. I came up with 50 pictures. So, clearly the issues lies in the underlying query used to produce the chart data. I have not yet had time to go back and figure out where the error is, but this pretty much isolates it to that query. It is clearly not in the real data!

Thanks again to everyone who acted as a sounding board for me on this!

Jack

One other possibility is the charting software itself. Could the X-axis (time) be assuming 100 minutes between hours? That would produce a "band" covering 60 minutes and nothing in the final 40 minutes of an assumed 100 minute period. I see there are absolutely NO data points in the last approximately 40% of each "hour."

_________________
Bryan Kinkel
BSK Consulting
Specializing in small land management