The Future of Urban Greening: A Study of One Million Trees to Determine the Best Species to Plant

The many greening projects planned by cities face challenging issues, including which tree species should be planted today to resist the climate of tomorrow. We wanted to supplement existing tools and catalogues by studying the species in different territories with respect to their resistance to drought stress. To do this, we studied the behavior of a million urban trees, corresponding to over 600 different species, during the last three summers under contrasting conditions. Here is our initial feedback on this study.

The framework of the study: resistance to drougth stress

Like any living organism, trees have a vital need for water. See how this golden raintree sees its cooling capacity regain its vigor the day after watering, whereas the day before, it had ceased all evapotranspiration.

fig.0 - Impact of water on tree behavior.

The drought we are currently experiencing raises fears of serious consequences for the spring and summer seasons, as climatic conditions are likely to become the standard regime for years to come. Therefore, it is essential to identify which tree species have the greatest potential for survival, which suffer the least from drought stress.

Various resources are available to help communities and all greening stakeholders to choose suitable species (such as SESAME   or Canographia  for France). For our concern, we were interested in the drought stress of trees and their ability to resist.

This study started last year, with the analysis of tree drought stress in different urban parks.

By way of illustration, here is a comparison of drought stress of each tree on the tree patrimony of Bordeaux – France between July 2021 and July 2022.

fg.1 - Drought stress of the trees - Bordeaux, July 2021 and 2022

The stress is characterized from the NDWI2 index.

Two lines of development have been implemented. The first consisted of monitoring the drought stress of trees dynamically, as shown in this video which traces the evolution of drought stress in the tree heritage of Bordeaux Métropole.

The second consisted of applying the same method to the documented tree patrimony (having the location of the trees and at the very least the genus and species) over three summers with different characteristics:

• 2020 : warm without excessive drought (in general)
• 2021 : rather cool, with rare periods of heat and high humidity
• 2022 : very hot and dry

The data thus acquired should make it possible to identify behaviours and trends.

The data

We have worked on documented tree patrimony available in opendata on various cities or state sites.

The map below shows the sites concerned over a vast territory stretching from Brussels to Montpellier and from Bayonne to Strasbourg. This first sample makes it possible to cover various climates and typologies of spaces: oceanic, Mediterranean, continental, plain, mountain...

Eventually, our study covers just over a million urban trees, representing more than 800 species, distributed as follows

fig.2 - Area of the sudy

Not all assets (in their accessible version) have the same level of detail, but this could be completed later.

The parameters accessible in the study concern:

• The species, obviously
• The place of planting: streets, parks, gardens, etc.
• The nature of the soil
• The age
• The crown size
• ….

Multiparametric analysis is therefore very broad and rich in information. We present here only a few examples of use.

For each city, a map is created for each year:

fig.3 - Tree maps of Strasbourg and Lyon

Some results and tools

All the data is accessible with a tool allowing to choose the site, species and period of analysis.

Take the case of the tree patrimony of the Eurometropole de Strasbourg, with 80,000 trees. Here is the population distribution of 4 species according to the level of drought stress at the end of July 2021:

fig.4 - Ratio of population for 4  tree species depending on their stress level

It clearly highlights different behaviors: Black elder has the lowest exposure, unlike the hop hornbeam, whose population is centered on a higher level of stress. However, we must go further. Early mapping analysis clearly indicated, unsurprisingly, that isolated street trees suffered more than trees in groups or in parks. The distribution of the average stress of the 636 species present in Bordeaux Métropole, according to the rate of "street trees," shows the same trend:

fig.5 - Stress level depending VS ratio of street trees

The distinction must therefore be made at the very least between street trees and other configurations.

If we are still interested in the tree patrimony of Bordeaux-Métropole, here are the population distributions of 4 species according to street trees VS others:

fig.6 - Ratio of population for 4 tree species depending on their stress level : parc (up) and street (down)

fig. 7 - Comparison of the tree population with respect to the stress level and street/other repartition between 2021 and 2022 

Overall, the distribution of the population by level of stress shifts sharply between 2021 and 2022, regardless of the territories. Opposite, the case of Bordeaux Métropole.

This shift is much more obvious on “street trees” than on “parks and gardens trees”.

The last two graphs show the details of the evolution of population distribution for 4 species between 2021 and 2022. Here again, the species mainly present in the streets suffer more than those in other areas.

fig.8 - Comparison of stress level for the 4 species : 2021 VS 2022

The shift is higher for Acer buergerianum, that is mainly located in streets.

Conclusion and perspectives

"This study has led to the creation of a substantial database, comprising up to 27 parameters per tree for a million trees over three years, which we plan to continue to enrich. Any suggestions in this direction would be welcome!

This database, combined with other data available on trees (such as Sesame) and local monitoring of species, can provide support for urban greening projects by proposing the most suitable tree species for each site.