Olivier Dambezat and Philippe Trouillet, Ceiba
When a tree falls in a public, visible and frequented area, the managers’ reaction is usually swift. Whether the failure caused damage or not, emotions run high and orders to diagnose the safety of nearby trees are immediate. On everyone’s lips, a single question: will they fall too?
De sedibus et causis morborum per anatomen indagatis ‘Of the seats and causes of diseases investigated through anatomy’ by Giovanni Battista Morgagni, first published in 1761.
In most cases, before any other intervention, the tree is cut down and its remains removed as quickly as possible. This rush is driven by practical and understandable concerns about freeing up public space, psychological concerns about calming fears, or, perhaps, political worries about judgement regarding the local management of trees. As a result, the investigation into the causes of the failure, which could guide the assessment of the mechanical condition of the remaining trees, will begin with a serious handicap: the lack of a corpse. And yet, both crime fiction and medical literature teach us that autopsies (from the Greek autopsia, meaning to see for oneself) are a formidable source of evidence and answers.
Without falling into anthropomorphism by comparing trees to people, the history of Western human medicine can be a source of inspiration for anyone interested in the development of tree diagnosis disciplines.
Until the 18th century, medical thought was dominated by a classification-based understanding of diseases. There were species of disease, just as there were species of animal or plant, and diseases had an autonomous existence, independent of the patient and his or her body. It was thanks to the systematic practice of autopsies and comparative work by a number of physicians at the time that the status of disease gradually changed. A new idea was developed that disease leaves traces in the tissues, and that it is these observable lesions that create the symptoms of the disease. This was the birth of a medical discipline that is still practised today: anatomical pathology. Morgagni, an 18th-century Italian physician considered to be the pioneer of modern anatomical pathology, believed that anyone who ‘has dissected or examined many corpses has at least learned to doubt. The others, who know nothing about anatomy and don’t bother to look into it, don’t have the slightest doubt.’
In those days, medicine was learnt at university and from books, a theoretical knowledge that limited contact with patients to the strict minimum. Life expectancy was 24.7 years (the figure for France in 1740).
Illustration of a hypothetical damage pattern on a Second World War bomber. Based on an unillustrated report by Abraham Wald (1943).
Post-failure diagnosis of a plane tree in Var, France, April 2022.
You are missing what is missing
Why should we pay attention to fallen trees? Isn’t it more interesting to examine those which remain standing?
As early as the 17th century, William Harvey, an innovative English physician, wrote that ‘The examination of the corpse of a single man who has died of a chronic disease is more useful to medicine than the dissection of ten hanged men.’ Anatomical pathology has made it possible to identify the traces of disease within the body, and each anatomical alteration corresponds to a functional alteration. After death, the pathological changes remain visible and can be observed, described and analysed.
By making it possible to think calmly and methodically about the visible symptoms, work on recently failed trees helps to mitigate the influence of numerous biases when diagnosing standing trees: anchoring bias (remaining focused on the first impression), pessimism bias and defensive arboriculture (‘you never know’), etc. But the main bias that post-failure diagnosis tackles is survivor bias. This bias consists of over-estimating a phenomenon by focusing on successful cases and ignoring all the others.
A famous example comes from the Second World War, when surviving aircraft returned from missions. After statistically analysing the location of the impacts of enemy fire on the fairings, American engineers proposed adding shielding to the areas most affected (for weight reasons it was not possible to shield the whole aircraft). But a mathematician, Abraham Wald, stopped them, arguing that if these planes had returned to their home base, it meant that they could continue to fly when hit at these points. Those that had been hit in other places were not included in the statistics because they had disappeared. The areas least affected on the surviving aircraft were therefore probably the most critical. He recommended shielding those areas, which was done. Without assuming that the war was won thanks to Wald, we can nevertheless see here that focusing on the survivors can lead to real diagnosis errors: by overlooking what is not there, ‘you are missing what is missing’.
An opportunity to learn
In arboriculture, as in military engineering, being able to analyse only intact subjects will not allow us to understand what probably differentiated the one that fell. Every tree failure should be an opportunity to learn.
However, in a post-failure diagnosis, two conditions are necessary for this observation experience to become a learning experience. The first step is to understand the symptomatology, which requires a number of parameters to be systematically and factually recorded. At this stage, the most common error is again a type of survivor bias. Since the absence of a symptom is itself a symptom, it would be a serious mistake to overlook, for example, the lack of adaptations (reinforcing wound wood, adaptative growth, buttresses, etc.) in the failed subject. The second condition is that we are dealing with a predictable failure that could have been anticipated. As in medicine, a certain number of failures are in fact associated with phenomena that have no visible symptoms (or at least, as these symptoms are not yet known or identified, they can’t be detected). In these cases, we probably can’t draw much from our observations, other than to confirm this occasional unpredictability.
In the field of life sciences, and particularly in arboriculture, there are undeniable difficulties in meeting the conditions for the development of real expertise as identified by the psychologists Kahneman and Klein: 1. A stable environment (regular, always behaving in the same way); and 2. Sustainable practice with quick, unambiguous feedback. If it doesn’t solve everything, the post-failure diagnosis could compensate a little for the lack of quick feedback (trees are a long-term process) by making the situation clear: there’s no ambiguity – it’s fallen.
The practice of post-failure diagnosis
For several years now, we have been systematically studying the failed subjects we encounter, sometimes by chance, sometimes following instructions from managers. The remains of the trees are cut up, and moved if necessary, for a meticulous process of observation, probing and sometimes dissection that can take several hours, until the phenomenon is understood. In the vast majority of cases, we are able to associate failure with a clear symptomatology. Once the symptoms have been identified, the rest of the diagnostic investigation is usually a simple matter: find any similarities in the surrounding subjects that are still standing and whose situation is of concern.
In nine years of activity and several dozen post-failure diagnoses, we have come across a few cases that could probably not have been anticipated (gale force winds, summer branch drop, etc.). These exceptional cases do not generally call into question the mechanical resistance of neighbouring trees. The vast majority of the trees studied showed a combination of symptoms, more or less obvious or easy to observe, which clearly explained the failure.
A final comment concerns the time lag between tree failure and post-failure diagnosis. As Morgagni pointed out in his time, an autopsy is of no use if we are unable to detect changes that deviate from the norm and are not due to post-mortem decomposition. Even more than disease, death alters tissues, and trees that failed long ago will no longer have much to offer the diagnostician’s eye.
In conclusion, if our aim is to make an accurate diagnosis that will enable the preservation of trees, we could also draw inspiration from Laennec, whose anatomical pathology-fuelled conception of disease led to the invention of the stethoscope, a tool that would pave the way for what could be called the autopsy of the living.
Originally a climbing arborist, then a consultant and trainer, in 2015 Philippe Trouillet founded the Ceiba consultancy and training centre in France – www.ceiba-conseil.com. He now specialises in clinical mechanical assessment and tree risk assessment.
Olivier Dambezat is Ceiba’s CEO. After a degree in philosophy followed by an artistic career, he has been associated with Philippe since 2018, working on the development of various subjects in amenity arboriculture. Philippe and Olivier will be presenting on ‘Mechanical features: towards new methods of building an observation-based likelihood of failure’ at the Association’s conference in September.
References
Drénou, C. (2022). Slow Tree : apprendre à respecter la lenteur des arbres. La lettre de l’arboriculture 106. May/June, SFA.
Foucault, M. (1983). Naissance de la clinique. Une archéologie du regard médical. Presses universitaires de France (PUF).
Grmek, M.D. (dir.) and Risse, G.B. (1997). La synthèse entre l’anatomie et la clinique. In: Histoire de la pensée médicale en Occident, Vol.2: De la Renaissance aux Lumières. Seuil.
Kahneman, D., and Klein G. (2009). Conditions for intuitive expertise: A failure to disagree. American Psychologist 64(6): 515–526.
Laboisse, C. (2020). Laennec, formation et destin d’un médecin, de l’anatomie pathologique au stéthoscope. mediaserver.univ-nantes.fr (accessed 20/3/24).
McRaney, D. (2015). Missing what’s missing: How survivorship bias skews our perception. www.ted.com/watch/tedx-talks
Tremblay, C. (2015). Compte-rendu: Naissance de la clinique de Michel Foucault. Aspects sociologiques 22: pp. 169-180.
Trouillet, P. (2022). Les biais de diagnostic. La lettre de l’arboriculture 105. March/April, SFA
Vallin, J. (1989). La mortalité en Europe de 1720 à 1914: tendances à long terme et changements de structure par sexe et par âge. In: Annales de démographie historique. Le déclin de la mortalité: 31–54.
Wald, A. (1943, 1980). Reprint of A Method of Estimating Plane Vulnerability Based on Damage of Survivors. www.semanticscholar.org (accessed 20/3/24).
This article was taken from Issue 205 Summer 2024 of the ARB Magazine, which is available to view free to members by simply logging in to the website and viewing your profile area.