Hellickson TrailCam Survey Questions

[grundsow]

August 2007 issue of Quality Whitetails has an article by Mickey Hellickson titled �A View From Afar� where he gives details about how to survey a deer herd with trailcams. He calculated buck-doe ratio, deer density, fawn recruitment and more for a 1,480-acre (2.3 sq. mi.) tract in Iowa, by using strictly the pics from strategically placed trailcams.

He recommends:
*baited cam sites
*1 cam for every 100-160 acres
*survey period of only 10-14 days (claims it captures 90%-100% of unique buck)
*survey either pre-hunt (Aug.-Sept.), or post-hunt (after last day of season yet before antler drop)

QUESTIONS:

Why are these parameters important?

I have 3 cams in a woodlot of maybe 50 acres that�s surrounded by farmland, orchard, and other woodlots. Before I start going back thru my pics and counting, is my setup sufficient for survey purposes?

Is bait really required? It�s not much of an option around here.

Must I spread my cams over more land mass, or can I just increase the time period of my survey to a point where I stop spotting new buck?

Hellickson�s deer density calculation is rooted in the number of individually identifiable buck captured on the trailcam pics (and then simply multiplying by number of doe sightings per buck, and then that result by number of fawns sightings per doe). I realize Hellickson�s reputation, but how reliable is this method?

I know that buck sightings on MY cams are very sporadic. Some buck are seen once and never again. Some buck are seen multiple times on one cam but never seen on another nearby cam.

So, it seems I would get different results depending on which cam I made calculations by. And if I were to combine results of all 3 cams, I would get yet another result. Does bait reduce this variance, and therefore can higher cam density make up for lack of bait?

For deer density calculations, since there are more than 50-acres of agriculture adjacent to my 50-acre woodlot, do I assume my cams capture all deer within a 100-acre area (50+50=100)? Do I assume 150-acres? Do I assume 640-acres (an average home range size)?

Thanks

 

[BSK]

Mickey's article is just describing the standardized photo census technique Jacobson first developed in the late 90s.

Yes, it works. But what pre- and post-season baited photo censuses tell you is who is using the property at those exact times. Deer can shift around between pre- and post-season, plus you have quite a bit of temporary range shifting and expanding that goes on during the rut. Bucks that never use your property except during the rut would not be picked up on pre- or post-season censuses, but they are still "manageable" (harvestable) bucks.

In addition to pre- and post-season baited censuses, I also like to run full-season unbaited censuses. The big difference is in camera density and placement. In a baited censuses, cameras are set up on a grid, generally 1 per 100 to 160 acres and bait is placed in front of the cameras to draw the deer to the cameras for photographing and censuses. In an unbaited censuses, cameras are place along trails, or at food sources (or my personal favorite--over both natural and mock scrapes) to census the herd. The deer are not "drawn" to the cameras, so an unbaited census requires more cameras, more camera locations and more time. But a season-long unbaited census WILL catch those seasonal range-shifters.

To answer some of your specific questions:


Why are these parameters important?

Because those are the parameters used in the original Jacobson study, and are time-proven to work well.


I have 3 cams in a woodlot of maybe 50 acres that�s surrounded by farmland, orchard, and other woodlots. Before I start going back thru my pics and counting, is my setup sufficient for survey purposes?

3 cameras on 50 acres is far more than enough, but to get a good unbaited census, you need to use as many different types of camera set-ups as possible and move your cameras frequently. If cameras are placed on one type of set-up, you can get highly skewed numbers. For instance, cameras pointed into food plots often get far more doe and fawn pictures than buck pictures, even in a balanced herd. Cameras placed over scrapes generally get more buck pictures than doe pictures, etc. You need to mix up locations as much as possible.

Some deer do not seem to be bothered at all by trail-cameras, but other deer--especially many mature deer--do not like tham at all, and will very quickly learn to avoid camera placements, hence the need to move cameras frequently.


Is bait really required? It�s not much of an option around here.

For a baited census, yes, but for an unbaited census, no.


Must I spread my cams over more land mass, or can I just increase the time period of my survey to a point where I stop spotting new buck?

If bait is not used, it is CRITICAL to move the cameras as frequently as possible. I prefer to move mine weekly if I can.


Hellickson�s deer density calculation is rooted in the number of individually identifiable buck captured on the trailcam pics (and then simply multiplying by number of doe sightings per buck, and then that result by number of fawns sightings per doe). I realize Hellickson�s reputation, but how reliable is this method?

The actual formula is:

TB = Total Bucks in photographs
TD = Total Does in photographs
TF = Total Fawns in photographs
UB = Unique, individual Bucks represented in all photographs (how many unique bucks do you have on film?)

To calculate the number of UD (Unique Does):
UD = (UB * (TD/TB))

To calculate the number of UF (Unique Fawns):
UF = (UB * (TF/TB))

The adult sex ratio (does per buck) is:
TD/TB

and the fawn recruitment rate is:
(TF/TD) * 100


These numbers are based on the concept that--on average--each unique deer is repeat photographed the same number of times. And that is where the data can get skewed away from reality. Overutilization of bait by one sex will skew a baited census. Over-utilization of one type of camera set-up in a non-baited census can skew the data.

However, what is by far the most important factor is the trends from year to year, not necessarily the exact numbers. Just ensure that once a census system is developed and implemented, that it is done the exact same way every year.


I know that buck sightings on MY cams are very sporadic. Some buck are seen once and never again. Some buck are seen multiple times on one cam but never seen on another nearby cam.

That's OK. It all comes out in the calculations. The calculations are based on the average number of times each buck is photographed.


So, it seems I would get different results depending on which cam I made calculations by.

That is correct.

And if I were to combine results of all 3 cams, I would get yet another result.

That is correct, and why it is so important to use multiple cameras. Some locations may be frequented more by bucks while other areas by does. The more cameras, the better the data.

Does bait reduce this variance,...

Not that i have seen.

...and therefore can higher cam density make up for lack of bait?

ABSOLUTELY. I get good unbaited season-long census data using 1 camera per 80 acres. You're already better than triple that. But again, I MOVE CAMERAS FREQUENTLY. From September through December I will shift 6 cameras through a grand total of 60-70 different locations (about 10-12 locations per camera).


For deer density calculations, since there are more than 50-acres of agriculture adjacent to my 50-acre woodlot, do I assume my cams capture all deer within a 100-acre area (50+50=100)? Do I assume 150-acres? Do I assume 640-acres (an average home range size)?

Contrary to the literature, you cannot get absolute accurate density numbers from square mile or less photo censuses. It would take a long discussion to explain why, but the statistics get very hairy (in fact, i still haven't perfected them yet). But even when censusing an entire square mile, the density numbers will be AT LEAST twice as high as the actual deer densisty. In essence, if the average deer density is 30 deer per square mile for the area, expect to census at least 60 deer in a square mile area. And that's why I don't put much faith in the exact density calculations, but instead use those numbers as "trend data" from year to year.

 

[grundsow]

In addition to pre- and post-season baited censuses, I also like to run full-season unbaited censuses.

Interesting!

Do the two methods produce similar results in terms of buck:doe ratio, fawn:doe ratio, buck age structure, etc? Or do seasonal patterns or something else cause significant differences?

In an unbaited censuses, cameras are place along trails, or at food sources (or my personal favorite�over both natural and mock scrapes) to census the herd.

You mean I can include scrape cam pics too? Wow! That will really boost my buck numbers.

I�m still amazed at how many different buck will show up at a single scrape, yet so many of them will never be seen on any other cams I have out in the immediate area. I often wonder how they find the scrape while coming in from all different directions.

3 cameras on 50 acres is far more than enough, but to get a good unbaited census, you need to use as many different types of camera set-ups as possible and move your cameras frequently.

Really? This is unexpected good news.

I keep my cams out 365 days a year and so could choose pics to count for any time period. I typically move the cams too every month or so (sooner if I get few pics). However there is one on an adjacent property that I�ve begun to keep over a salt lick permanently.

These numbers are based on the concept that--on average--each unique deer is repeat photographed the same number of times. And that is where the data can get skewed away from reality. Overutilization of bait by one sex will skew a baited census. Over-utilization of one type of camera set-up in a non-baited census can skew the data.

This was my main concern about the methodology.

Contrary to the literature, you cannot get absolute accurate density numbers from square mile or less photo censuses. But even when censusing an entire square mile, the density numbers will be AT LEAST twice as high as the actual deer densisty. In essence, if the average deer density is 30 deer per square mile for the area, expect to census at least 60 deer in a square mile area.

Hmmm, you must have anticipated my next post�

What you just wrote is just the opposite of what Hellickson says. He claims that his cam survey of over 2 sq. mi. was likely CONSERVATIVE due to not capturing all buck. So how does one reconcile these different points of view?

BTW, your post is a little more comforting , as I about fell out of my chair when I read Hellickson�s survey concluded a pre-hunt deer density of whopping 68dpsm! Is that really hunting when you can practically walk across their backs?

To further aggravate me, he counted 158 total deer pre-hunt and they harvested only 11 (or 7%) total deer!! He called their doe harvest of 10 a �stepping-up� of the antlerless harvest. How is that QDM???

I mean, he noted that recruitment amounted to 27 fawns/sq. mi. So, with their measly harvest of only 5 deer/sq. mi., the overwinter deer density would climb from 42 to 64 deer/sq. mi.!! Bleeech!

Even if you cut the numbers in half, they still left the herd GROW from 21 dpsm up to 30 dpsm.

 

[BSK]

In addition to pre- and post-season baited censuses, I also like to run full-season unbaited censuses.

Interesting!

Do the two methods produce similar results in terms of buck:doe ratio, fawn:doe ratio, buck age structure, etc? Or do seasonal patterns or something else cause significant differences?

Honestly, depending on how many years a baited census has been run consecutively, I think an unbaited census gets better buck:doe ratio numbers, AS LONG AS NUMEROUS DIFFERENT CAMERA SET-UPS ARE USED. As I mentioned, pointing a camera into a food plot is going to get numbers skewed towards does and fawns, while cameras over scrapes will get numbers skewed towards bucks.

One thing about baited censuses I've noticed, and others have commented on, is over time, older bucks tend to avoid the bait. By the third or fourth year a pre- or post-season census is run, mature bucks seem to "wise up" to the bait and won't use it, eliminating them from the census.

In addition, a pre-season baited census only gets the deer that are there at the time. Which deer are using a small property can change dramatically over the course of an entire season.

 

[BSK]

In an unbaited censuses, cameras are place along trails, or at food sources (or my personal favorite�over both natural and mock scrapes) to census the herd.

You mean I can include scrape cam pics too? Wow! That will really boost my buck numbers.

I�m still amazed at how many different buck will show up at a single scrape, yet so many of them will never be seen on any other cams I have out in the immediate area. I often wonder how they find the scrape while coming in from all different directions.

Scrapes are probably the best place to inventory bucks, as nearly every buck that wanders through the area will stop to investigate the scrape.

In many years, scrapes will be the only location where I pick up a particular buck. I may get 3 or 4 pictures of him during the season, but all of those pics will come off scrapes. If I didn't use scrapes as a camera location, I would never "capture" that buck in the inventory.

 

[BSK]

These numbers are based on the concept that--on average--each unique deer is repeat photographed the same number of times. And that is where the data can get skewed away from reality. Over-utilization of bait by one sex will skew a baited census. Over-utilization of one type of camera set-up in a non-baited census can skew the data.

This was my main concern about the methodology.

Again, that's why it's so important to stick with the same methodology from year to year, so that any "skewing" of the data is done the same way every time. This will maintain your trends.

I realize some just will never "get it," but when running any type of "monitoring" data system, the trends are far more important than the actual numbers. Too often hunters/managers get hung up on worrying about exactly accurate data, yet the trends being accurate is a far more important concern.

 

[BSK]

Contrary to the literature, you cannot get absolute accurate density numbers from square mile or less photo censuses. But even when censusing an entire square mile, the density numbers will be AT LEAST twice as high as the actual deer densisty. In essence, if the average deer density is 30 deer per square mile for the area, expect to census at least 60 deer in a square mile area.

Hmmm, you must have anticipated my next post�

What you just wrote is just the opposite of what Hellickson says. He claims that his cam survey of over 2 sq. mi. was likely CONSERVATIVE due to not capturing all buck. So how does one reconcile these different points of view?

Is his density calculations an under-estimate or his buck inventory an under-estimate? You may miss some bucks with any census technique.

The density numbers in a square mile census can't possibly be accurate unless the square mile is surrounded by a high-fence. Deer move around within their range, and many deer will have ranges of a square mile or more. If two neighbors each with a perfect square mile of land adjoining each other run a census, and neighbor "A" gets a picture of a given buck and neighbor "B" also gets a picture of that buck, who gets to count that buck in their density calculation? He can't be counted in both. Hence, when censusing a square mile area, the census taker must realize that nearly every deer they "count" would probably also be counted by a census taker on an adjoining square mile, hence the deer has to be "split" between square mile areas. In fact, a particular der might be captured in three or more square mile blocks, hence have to be split equal-ways between each square mile census. That's why it's a safe bet to split every deer counted in a square mile census in at least half. You assume that every deer crosses imaginary square mile boundaries. When I run a census, I immediately split the density numbers in half, knowing it probably needs to be split more ways than that.

The problem is, this has never been addressed in the scientific literature. Although when it discussed, all will admit this is the statistical reality of censusing. For example: assume we have a perfect 2-square-mile area under high fence--two adjoining perfect square miles side by side surrounded by a single fence. Run a baited census in the left-hand square mile. Then run a census in the right-hand square mile. I'll bet 95% of the deer inside the fence move between the two square-mile sections, hence would be picked up in both censuses. They can't be account for in both square mile areas. Let�s say there are 50 deer inside the 2-square-mile fence, making the actual density 25 deer per square mile. Yet a census in the left-hand square mile counts 45 deer (95% of all the deer inside the fence) while a photo census in the right-hand square mile also counts 45 deer (95% of the deer move back and forth across the imaginary square mile boundary). That would lead the census takers to assume there are 90 deer inside the fence, yet there are only 50.

 

[BSK]

BTW, your post is a little more comforting , as I about fell out of my chair when I read Hellickson�s survey concluded a pre-hunt deer density of whopping 68dpsm! Is that really hunting when you can practically walk across their backs?

You think that is frightening, you should see the numbers over a 6 month period of time. Remember, some deer will shifting around quite a bit over the course of a hunting season, especially when you consider the influence of the rut, and how bucks can either expand or shift their range while looking for receptive does. This can temporarily bring individual deer onto a small property where they will be camera censused. These temporary shifts and inclusion into the calculations will really add up in a season-long census/inventory. If adding the deer numbers cumulatively over 6-months, this can add up a huge number of deer, even though at any given point in time, nowhere near that number of deer is using the property.

For example, over a 6-moth period of time, last year my total cumulative camera census density numbers for my 488-acre property where 157 deer. Considering I'm photo-censusing 76.25% of a square mile, theoretically the deer density would be 206 deer per square mile. However, if you look at the camera data for any given two-week period of time, the numbers are nowhere near that high. The number only gets that high due to cumulative counting of deer that are moving onto and off the property during short (or not so short) periods of time. Once I account for the "sharing" of deer between adjoining square-mile sections, the deer density calculations from last year were just about where they should be--about 35 deer per square mile--if only two week's worth of data is analyzed--any two weeks you want to look at. The one difference would be around the rut, when a two-week calculation will come up closer to 40-45 deer per square mile, but that is to be expected, as deer are moving farther during the rut (larger ranges, hence more "sharing" of deer between adjoining square-mile sections).

However, there is one thing to remember about these over-estimated deer density numbers due to "sharing" of deer between square mile sections, and that is those numbers are "real" when it comes to manageable deer. If you have a square mile of land, and the deer density is really 35 deer per square mile, at least 70 deer are going to walk across your property during any two week period of time, MAKING ALL 70 OF THOSE DEER MANAGEABLE, HARVESTABLE DEER. Likewise, although last year my census calculations over 6 months came up with 157 deer while the deer density is only really around 35 deer per square mile, all of those 157 deer were manageable harvestable deer at some point during the hunting season, even if many of those deer only crossed the property a few times on a few days. In fact, over 6 months, we photo inventoried 47 unique bucks on the property, During any given two week period, no more than 20 were using the property at that time, but of the 47 unique bucks, 46 were photographed on the property some time during the legal hunting season, making all 46 manageable, harvestable bucks.

It is this very situation that makes small-land management so successful--because the small-land manager has he opportunity to "influence" many more deer than the deer density would suggest they can influence.

 

[BSK]

To further aggravate me, he counted 158 total deer pre-hunt and they harvested only 11 (or 7%) total deer!! He called their doe harvest of 10 a �stepping-up� of the antlerless harvest. How is that QDM???

I mean, he noted that recruitment amounted to 27 fawns/sq. mi. So, with their measly harvest of only 5 deer/sq. mi., the overwinter deer density would climb from 42 to 64 deer/sq. mi.!! Bleeech!

Even if you cut the numbers in half, they still left the herd GROW from 21 dpsm up to 30 dpsm.

And now you can see why hunters/managers can shoot and shoot and shoot does until the cows come home and have little impact on the local deer density IF THEIR NEIGHBORS ARE NOT SHOOTING DOES.

 

[Anonymous]

Some deer do not seem to be bothered at all by trail-cameras, but other deer--especially many mature deer--do not like tham at all, and will very quickly learn to avoid camera placements, hence the need to move cameras frequently.

i read an article one time by a game manager for a big farm in ga. that said he baited the deer for say a month and then put the cameras in for just a couple of days and claimed to get pics of like 95% of the bucks they would see or kill on that farm that year. does this seem plausible and a strategy to get around the situation you describe? thanks.

 

[BSK]

Some deer do not seem to be bothered at all by trail-cameras, but other deer--especially many mature deer--do not like tham at all, and will very quickly learn to avoid camera placements, hence the need to move cameras frequently.

i read an article one time by a game manager for a big farm in ga. that said he baited the deer for say a month and then put the cameras in for just a couple of days and claimed to get pics of like 95% of the bucks they would see or kill on that farm that year. does this seem plausible and a strategy to get around the situation you describe? thanks.

The scientific literature (the orginal study done on camera censusing) found that they got 96% of the deer in the area at that time on film in 14 days. This is with a pre-baited census, and yes that is physically possible. But when you start talking small properties (those 1,000 acres or less), seasonal range-shifting of deer can mean that deer not on the property during the baited census hence not censusable can be on the property during deer season. This is the most common reason hunters think they aren't getting all of the deer on camera during a pre-season (summer) baited census. They may actually get all the deer on the property at that time, but different deer can be on the property during deer season. They may see and kill bucks during the season that were not in the summer census, leading them to believe the census technique didn't work.

For example, in my previous response to grundsow I mentioned that last year we photo-inventoried 47 unique bucks on my property from August 1 to Jan 31. I first ran a baited census in August for two weeks and then shifted to an unbaited census from mid-August through the end of January. Of those 47 bucks, only 13 of them used my property during the baited census period and were inventoried. The other 34 didn't show up until after the baited census was over. And of those 34 that didn't show up until after the baited census was over, 10 were only photographed once, meaning they were probably just passing through or are were very rare visitors to the property, reducing their manageablity and harvestability. The other 24 were seasonal shifters that shifted ont my property only during the fall or only during just the peak of the rut.

On average, 50% of the bucks we kill each year are "residents" that were picked up in the summer census and the other 50% are seasonal range-shifters that don't show up until after antler velvet shedding.

 

[grundsow]

the other 50% are seasonal range-shifters that don't show up until after antler velvet shedding.

What does that mean?

I have one large buck (now an 11+ pointer with maybe a 22�+ spread) in particular that I have captured on cam for 3 years now, he must be 5+ years old at this point. The only daytime pics I have of him are one in velvet, and one post-shedding (I�m nearly certain it�s him). I�ve even located a rub line that�s maybe 400 yards long on trees 6� and even 9� diameters that I must attribute to him. All this on the same 50 acres or less even though I�ve also captured him 1.5 miles away too.

The strange part is that while he displays tarsal staining that runs down to his hooves, I�ve never ever caught him at a scrape in either location. Are you saying he may vacate these areas almost completely at certain times of year? Why wouldn�t he at least visit scrapes in these areas where he�s seen at other times of year?

Thanks BTW for all the detailed info, I'm not ignoring you, just need time to digest and respond.

 

[BSK]

Yes, you can have bucks that visit your property only during one season of the year or even one portion of one season. This is especially true during the rut. A buck may completely leave his normal range and go to a second area where he spends the peak of the rut, then returns to his normal range. These two ranges may have no overlap at all.

In the past, as much as 50% of the bucks using my property in summer would leave right at antler velvet shedding time, not to be seen again until the next summer. However, since greatly improving our habitat, fewer bucks leave after summer, generally just one or two. But in most years we get a big surge in "seasonal-shifters" that spend their summers "somewhere else" but move to my property right after antler velvet shedding time and generally spend the entire fall and early winter on my place, only to leave in spring.

These type patterns are very common no matter where the data comes from.

Although, as more and more detailed GPS-collar studies are done--where older bucks are followed on an hourly or even sub-hourly basis--we are finding there are about as many "patterns" to home ranges are there are older bucks. Each buck is an individual.