Thermal imaging its here to stay

[arb culture]

The point is that the accuracy of the data capture with thermal imaging, which provides information about the current physiological condition of the tree, can be used in statistical analysis of populations. This allows the calculation of a probability of failure within that population.

 

The point is with this is that it is possible to gather large amounts of data very quickly and inform far more holistic decisions that are based on accurate information about the tree in relation to its local population.

 

The probability of failure scoring that is provided with a thermal image assessment is based on the use of a computer model within the TTMS software used to calibrate the images (see below). This information can be used in other statistically analytical systems as well. It all comes down to the availability of data to analyse.

 

Thermal imaging allows the capture of information about tree populations that has never before been practical let alone affordable. In the right conditions we can survey up to 100 trees a day, capturing thermograms of each tree that will enable an assessment of its physiological function from roots up into the branches.

 

What other tool on the market can offer you this.....?

 

 

 

 

 

Visual Tree Assessment has two significant limitations. Firstly it is based on observations of the trees historic reactive growth process, because the arborists uses knowledge of how the tree grows and responds to structural weaknesses over time. Some VTA symptoms will take decades to develop as the tree slowly reacts to a particular structural situation or the presence of decay. Therefore VTA is an assessment of a historic process, and does not necessarily tell you anything about what the trees current condition is.

 

For example a VTA symptom such as an area of wound wood or reactive growth that has formed a bulge may now be dead wood, dried out or even decayed. The arborists is not likely to be able to say one way or the other, and this is really the second issue with VTA it is down to the inspectors personal opinion and experience. The inspecting arborist has to place a significance on the observations made, and then determine a conclusion without any supporting evidence.

 

The use of thermal imaging provide the inspector with an ability to determine what parts of the tree are still functional. Using heat as the tool it is possible to use the infrared camera to observe the water presence and movement within the tree. This provides the supporting evidence to demonstrate that the reactive growth associated with VTA symptoms is still functional. This is done simply by using the camera as a tool to observe heat movement within the tree.

 

The camera technology allows the capture of a fully radiometric digital thermogram, which includes all the temperature information that the infrared censing hardware can absorb. For example our camera is sensitive to a temperature range between -40 to +120 degrees Celsius. The infrared software that comes with the camera allows you to adjust the colour application to different temperature ranges and differentials. This is what create the colour image that we see. For trees the image temperature span (distance between upper and lower temperatures) tends to be no more than 2-3 degrees Celsius. The level (position of the set span on the sensitivity range e.g. -40 to +120 degrees Celsius) will vary depending on the time of image capture and the temperature ranges of the day.

 

The analytical capabilities of the radiometric thermograms means that it is possible to use this temperature data in a number of ways. This is where the post image capture processing comes in. We use TTMS (Tree Thermal Matrix System) software. This enables the comparison of the temperature data captured on the day with a computer model that predicts what the tree's thermal signature should look like (i.e. what its temperature should be) if it is in optimal condition.

 

The simple software calculation uses temperature data from the site and metoffice weather data to work out the temperature curve for the day of image capture. The product of the calculation is an upper and lower temperature range that allows the thermograms to be adjusted. The processed thermal image then provides a visual illustration of where the tree's thermal signature is.

 

If a tree is healthy then the processed thermogram will be very close if not identical to the actual temperature readings on the tree at the time of image capture. However, if the tree is not responding well to the temperature changes on a day to day basis it will be a degree or so cooler than the optimal. Such trees do not have sufficient functional wood to allow them to react to the daily temperature changes.

 

If a tree cannot respond to daily temperature fluctuations it is in poor health and also unable to grow reactively to the presence of growth stresses such as wounding, decay, drought, pruning or root disturbance. This then allows the arborist to make informed decisions about the trees management.

 

If significant areas of dysfunction are found, such as cavities or decay, it is then possible to work out whether the tree can respond to the challenge of living with these weaknesses or not.

 

 

What other tool on the market can offer you this.....?

 

 

 

 

 

In the image I posted earlier, this was of a drought stressed tree and there were some specific issues that had significantly contributed towards this. Not least the fact that roots had been left exposed following an investigation to determine damage to the adjacent wall. This in combination with the demand for water in the spring leaf growth period meant that the tree was suffering. The concentration of cooling associated with this presents a very symptomatic image and a recommendation to re-assess the tree when the wholes had been filled back in and the tree had benefited from a period of rain or been watered.

 

Seasonal or periodic drought stress is something that trees have to cope with on a regular basis, but it does expose them to the progress of decay fungi within their trunk and central root system. Where decay fungi are present (in 99.9% of mature trees) the period of drought stress provides them with a new food source that they can capitalise on. The fungi will take advantage of this opportunity.

 

In such situations one of the recommendations to the client is to water the tree, and reduce the amount of time it is exposed to drought stress. This will then limit the progress of decay.

 

 

 

What other tool on the market can offer you this.....?

 

 

 

 

Welcome to the future........

 

 

This is a new language in true tree care.

 

That's a very long answer.

 

I'm afraid I got a bit bored by the regurgitated stuff about the magic maths and camera sensitivity. I've heard it all before, many times.

 

So just tell me what you say when a client says, "Is that tree safe then?"

[arb culture]

Ah, Batiarb, I've just read your posts about auras etc.

 

Perhaps you could use your horoscope to predict whether or not a tree will fail.

 

Back to the science though, as I've said many times, I can see the value of FCIR imaging in tree assesment.

 

It might also be true that Marcus's TI system is the bees knees, but without it being open to scrutiny there is no way of knowing.

 

Marcus you say that TTMS is the ultimate in auditable data, how can that possibly be true when you can't show the client, the planner, the appeal officer etc how the probability of failure has been worked out?

Edited July 1, 2010 by arb culture
typo

[ben lightfoot]

Thank you for the reply i have been on at jim for months to send me some information after i found out that you were meeting up and he has sent me nothing. Im still at the sharp end of the industry and am just finishing my tech cert where we looked at some decay detection but not in much detail hence why i was interesed in the findings. After reading this post im starting to understand thermal imaging a lot more i personally want to see it used more for identifying wildlife and habitats within a tree especially ivy covered trees. I look forward to reading more articles.

 

Ben

[Marcus B-T]

 

Marcus you say that TTMS is the ultimate in auditable data, how can that possibly be true when you can't show the client, the planner, the appeal officer etc how the probability of failure has been worked out?

 

I have give three seminar series, one international paper, and some postings on this site about how the likelihood is calcualted. Everyone who gets a report from me has an explaination attached, and I recommend that every one who uses the sofware does the same.

 

I can only give a summary at the moment (just too busy) but basically you look at incidences of the amount, location and severity of dysfunction in the population and look at the extremities of these occurances in realtion to the dimensions of the tree. This data is then cross referenced against tree failure information and the mechanistic models. So the data is cross calibarted in three ways (highly robust) however the first method (the population analysis) is the most robust but has do be carried out for each town and city (hence the hiegt to dbh calculator). However, there is a critical amount of dysfunction at 70% for undisturbed trees and this can be used as an initial bench mark.

[Marcus B-T]

THe other thing is that I did actually post that there is a demo version of the sofware that you can arrange to have and you can orgaise a demo of the camera from the fanufacurers to go with it (try before you buy if you like). If that is not transparent, what is?

[arb culture]

THe other thing is that I did actually post that there is a demo version of the sofware that you can arrange to have and you can orgaise a demo of the camera from the fanufacurers to go with it (try before you buy if you like). If that is not transparent, what is?

 

Hello Marcus,

 

I have been following the development of TI for a while because it's interesting, because I see the value of the hardware, and because I want to see the value of the software.

 

So far I have not been able to access any of the calulations used in determining the stability of a tree, so I can't determine the value of the software.

 

Does the 'try before you buy' option allow access to the calculations used?

 

[arb culture]

I have give three seminar series, one international paper, and some postings on this site about how the likelihood is calcualted. Everyone who gets a report from me has an explaination attached, and I recommend that every one who uses the sofware does the same.

 

I can only give a summary at the moment (just too busy) but basically you look at incidences of the amount, location and severity of dysfunction in the population and look at the extremities of these occurances in realtion to the dimensions of the tree. This data is then cross referenced against tree failure information and the mechanistic models. So the data is cross calibarted in three ways (highly robust) however the first method (the population analysis) is the most robust but has do be carried out for each town and city (hence the hiegt to dbh calculator). However, there is a critical amount of dysfunction at 70% for undisturbed trees and this can be used as an initial bench mark.

 

Is that 70% of circumference, cross sectional area, transverse sectional area...?

 

How did you come to this 70% figure?

 

What about a tree with 70% dysfunction but with large amounts of solid (but non transporting) wood?

 

How do you calculate the differences in stability due to crown size and condition?

 

What about differences due to species?

 

Are different soil types and exposure taken into account?

 

What about hollow trees?

 

Etc...

 

I know you are busy, so are my questions answered in your peer reviewed papers, and if so, where can I get them?

 

[Amelanchier]

How did you come to this 70% figure?

 

It seems to be consistent with Mattheck et al and the 1/3 rule.

[Yorkshireman]

 

So just tell me what you say when a client says, "Is that tree safe then?"

 

Well if you take the data correctly & the data is then put through Marcus's system

& he gives you a report that’s says its safe or not safe , then if he's wrong you can tell the client & the judge

 

NOT ME GUV its him ,, he said so .,...

 

the math bah , Marcus isn’t going to give that one away is he !

He says he can defend it & has statistics that back the claim

 

So why worry if you've collected the data correctly & conveyed the report to the client

Nb this isnt designed to replace the arborist on site, merely to enhance them

& aid the rate of data collection & insure that the data can be verified later & this removes some of the subjectiveness

 

However Still have your eyes open , take a hammer & a probe

 

T I has limitations however it also brings alot to the party too

 

 

We recently did 63 200yr + limes & they had lots of issues that go along with that kind of seniority of trees don’t last for ever , however these will last longer than they may have done had the functionality not been verified

None to be removed 'Which is good ' just some minor corrective works

Edited July 5, 2010 by Yorkshireman

[Marcus B-T]

For copies of papers e-mail [email protected]

 

 

 

The TTMS process.

The TTMS system is essentially a statistical tool and compares individuals to large populations, the larger the better. It is also a decision tree and works as follows.

Initial site inspection

The initial part of the process is a site inspection to zone the potential risk if a tree was to fail in a given area. The conditions for this are in my recent paper in the Arboricultural Journal but they are not dissimilar to the methods proposed as part of THREATS and some other systems. The zoning also determines the frequency of re-inspection. The site is also given a stress index based on the growth of the trees within the wider population, and where stress is a consideration the re-inspections may be more frequent depending on the overall risk. For example in a Low Risk Zone re-inspection is less influenced by the stress trees are under than in a higher risk zone. The stress index is unpublished at the moment but will be published this winter.

Tree inspection

The trees are then inspected. Outcomes are then compared to the wider population data, e.g. if a tree has a bleeding canker, this may be present on one or two trees of a given species and may therefore be significant. However, it may be that most of the trees have the cankers and if most of the trees are relatively healthy, and there are not large numbers of trees falling down then although the cankers may still be noteworthy and a consideration with re-inspection, they are probably not a significant factor in risk.

Determining the need for further evaluation of attributes associated with potential risk.

Some attributes will be significant aspects of risk and will warrant further inspection, some factors will be significant aspects of risk and will not require further inspection. The later can come about for two main reasons; firstly the attribute that contributes to the risk is so blatant it would be a waste of money to look at it further, i.e. it can be evaluated visually; secondly the overall risk is still too low to warrant further inspection, i.e. the risk is apparent but remedial work will not significantly alter the overall risk. The decision as to whether further information is required will also be affected by the presence of a TPO or conservation area and may not be purely down to risk alone.

The visual survey will highlight trees that may require work, with the obvious trees the work requirement is almost certain and a further investigation will not change this, but with some trees the significance of the attribute identified as the cause may be less well established. At this point further investigation is required. The visual part of the risk evaluation can be seen as the attribute identification, this next part is the attribute evaluation.

Attribute evaluation

Not all trees require this next stage but if at tree is identified as having an attribute that may lead to failure at some point in the near future although it is not imminent then it is likely that it will become necessary to have the attribute evaluated.

The reason the evaluation is necessary is that when attributes are evaluated only 1% of them will lead to a failure within the next three years, so in most cases work can be scheduled in according to budget requirements, also around 75% will actually pose very little risk. What is the evidence for this? The low number of tree failures we see every year. If every attribute we look at leads to a failure we would see in the region of 40% of our trees failing in some way. So evaluation and monitoring will reduce costs but also allow for the trees to adapt to the attribute.

A good way of looking at this is that if the failure is not imminent but work is thought necessary then a further evaluation is probably warranted. The further evaluation will probably reduce or even remove the need for work in most cases, and in a rare number of cases it will highlight that more work is necessary.

What kind of evaluations can take place? The first is the evaluation of the amount of disruption an attribute causes. This can be looked upon as the amount, extent and location of dysfunction. When you study a population as a whole there is critical pathway of dysfunction development to a point of failure. That is dysfunction can develop in a region of a tree and as it does the tree stability is not compromised. Because of this there are large numbers of incidences of dysfunction within the population. As the dysfunction becomes more pronounced then there are fewer incidences present because either the tree is repairing itself or the attribute is lost in a failure. This can be cross referenced with actual tree failures and the amount of functional wood that has been lost. This was the subject of the seminars I gave and the paper at the ISA conference, Again it will be published in the winter. TTMS uses thermal imaging for this part and the system calibrates images against Tmax and a Tmin as described in my other paper in the Arboricultural Journal.

This assessment on its own is not enough to evaluate the attribute; it must be looked at in consideration of the overall structure of the tree and the tree’s location, since this is what ultimately causes the failure.

TTMS uses the population for this as well. This part is handled initially by the height to dbh calculator. Exposure is one aspect, has the tree been recently ‘un-sheltered’ by the removal of other trees or buildings? Another is the trees adaptation to it’s environment; does the tree fall within the height and dbh of the population? If it does it is likely to be stable and not require work. The height to dbh calculator is critical because the most common work advocated is crown reduction. The height to dbh calculator can also be used to back calculate an estimate of the amount of exposure. The workings of this is described in another current thread and there is a down load available.

I also have a calculator that tells me when there have been enough dry days to start to look for signs of sudden limb failure based on previous incidences.

The remit I have from a number of sources is to remove subjectivity form the process wherever possible. This is a continuing process.