Blog

More About Models

Move over Bella & Gigi Hadid, there is a new super-model in town!
We speak to our very own Data Scientist, Marco van der Heijden about “FastTrack”, the carbon-capture calculation model living on a diet of data, beautiful and elegant enough to take over all of the runways. We also find out more about his work and role in our tech team, and how he sees the future of technology as a tool for reforestation.

Could you tell me a little bit about your background, your role and how you got into working at Land Life Company?

My background is in physics which I studied at Delft. I have worked with writing algorithms and developing models in a range of tech-driven roles over the years. During the lockdown I was looking for a nice project to do, I was a bit done with doing work that I wasn’t inspired by and that was when I heard about Land Life Company. I find the work really valuable and interesting. I have many roles but the main one is defined as a data scientist. That consists of multiple things, looking for high-quality data and preparing it so it can be inserted into models. Also, I need to think about how we design our data warehouse, how to make better predictions using the models we work with and finally create models using the data we collect. 

Which is the main model are you currently working on?

They can be categorized in multiple ways, the main model I work on is the FastTrack model developed by our colleague Koen, (Professor Koen Kramer). This is a carbon capture prediction model which is more statistical. Other models are the Climate Growth Response model, which looks at how certain trees respond to the surrounding climate and deep learning models that are more related to drones and remote sensing.

Amazing, how does FastTrack actually work?

FastTrack is a statistical model that requires specific plant parameters specific to species and also to the area. For example, in Northern Spain, everything is super green and there is a lot of rain but in the south, it is completely different. So the trees also respond differently and the soil is different. So, one of the big challenges is to find the specific plant parameters that are tied to our target planting site, because in the end, we want to predict the specific carbon capture and how much mortality we can expect. When FastTrack receives the right input parameters, it models the growth of trees over a certain period and how certain species interact with each other in certain conditions.

What kind of data do you need to feed the FastTrack model?

For FastTrack we need plant parameter data, for example, how big can a tree get, how fast can it grow and the density of the wood. This data can be found in sources, for example, papers or public databases like National Forest Inventories. As our planting sites get older, we can also use our own data. Also, we can go into the field to perform Site Productivity Assessments. These give great insights, however, they can be time-consuming.

Do you use existing data or does Land Life Company collect data as well?

We use both. One of the most important data sources is the National Forest Inventory which is all over the world. It is kept up to date and is very helpful. This keeps track of how forests are growing, they go to specific areas and then return 5-10 years later to check again and monitor tree mortality, diameter, heights, how big the sample area was and general health – these are really important measurements. We also use climate data and look at soil type, there are also agricultural maps of that and finally, a lot of satellite data. More and more satellites are being released with this technology and different algorithms, these are quite amazing! These are all public datasets, then we have our private datasets which are growing every day. We go to the field and accurately measure trees with a Trimble to monitor how they are growing, what their status is and then if we keep returning we can get an accurate dataset regarding mortality and our carbon predictions. So this will be cool to see. Through all this, we can see which trees to plant where and this is the cycle, using the models as well means we can optimize and improve our plantings. 

Can you give us an example of this in practice?

One example could be the assessment we did last year, where we went to the field and we wanted to collect data from the field in the Burgos area. So in the end, based on the tree and plot data and the dendrochronology and soil samples, etc we took those to and made a site productivity assessment. These assessment results were compared to our own model results in terms of annual carbon capture and saw that the predictions were quite close to each other which gives us more confidence that what we are modeling is actually representing reality.

What do you do with the data produced by the FastTrack?

We use the FastTrack output to gain more insight on the potential carbon capture. We can tweak the input parameters like tree densities etc to see how these will affect the eventual capture versus planting costs. Although it must be carefully noted that costs and total carbon capture are not the only driving factors for the end result. For example, we always strive for biodiversity.

Why is tech such an important part of what we do at Land Life Company?

Tech is not only monitoring, it is also our planting technology and prediction technology. We can also use it if you look at climate growth relationships, we can use it in the future to prepare for extreme events that could happen. We have a very efficient way of monitoring, keeping track of what we grow and how we grow. Trees also produce more trees, not only the ones we planted. These are really important to study because we can take into account our carbon capture and follow biodiversity regeneration. It’s fun to think 40 years from now what will be there. 

Any final words or tech-driven wisdom?

We want to make forests more robust for climate change. It is amazing what data collection can do. Apart from the public data from our plantings, we are getting more and more information and this keeps on expanding. We are finding better ways to gather knowledge through our plantings, our predictions and our monitoring. I’m really excited to see what we can gather and this is just the beginning.