3D Tree - An Alternative Visualization Approach for Complex Data

Blender provides an Addon called “Sapling Tree Gen” which helps to create individual trees. The tool uses Bezier curves to build the tree. Starting from the main stem branching out until the desired level is reach. While the default sapling does not have a huge resemblance with the trees in Pfynwald there are a lot of parameters which one can tinker with to achieve a higher similarity. Furthermore, the bezier curves provide the basement for our flux visualization. On the one hand it is possible to use the curves itself to simulate the internal fluxes (as seen in Fig 3) and on the other hand we were able to use the control points of the curve as location for the atmosphere - tree interaction. 

Once the tree’s structure is established, the subsequent step involves incorporating real-world data. For this prototype, we concentrated on the VPD measurements taken in the VPdrought project. The measurement setup includes sensors placed above, within, and below the crown.

We interpolated the value for its position at a specific timestamp using at least the three closest neighbors to a branch tip.

Initially, we experimented with spawning molecules at a certain interval based on some measurement values. However, we soon realized that a basic simulation of the process within a leaf could yield more satisfactory results.

We want to emphasize strongly at this point that this simulation is solely for visualization purposes and does not aim to generate numerically accurate quantities of any flux or product.

Each tree’s end node is assigned a Python class that contains the previously added real-world data and the simulation.

The real-world data comprises the interpolated VPD readings and the measurement times.

While leaf metabolism is highly complex and depends on a vast array of factors, temperature, light, and water availability are critical to the ongoing processes.

We used the VPD as the primary driver of stomatal closure. VPD, or Vapour Pressure Deficit, describes the air’s water vapor saturation based on its temperature.

The measurement times are utilized to implement a day/night cycle.

For each timeframe, we begin by reading the VPD values and calculating the stomata’s opening level.

As the stomata regulate the flux between the atmosphere and the leaf, we can then simulate the storage of each element in the leaf, run all processes, and thus simulate new storage. This leads to either a surplus or a deficit of the elements, which then triggers the atmospheric exchange.

A significant challenge was to simulate the “decision” of the Rubilose biphosphate on which molecule it should bind and whether there will be photosynthesis or photorespiration.

To introduce some uncertainty into the decision, we added 10 cores, similar to a computer’s CPU cores, during the day and 2 during the night. Each core has the option to choose which process it wants to run during the day. The decision is based on a normally distributed random value with a mean of the current VPD value and a standard deviation of 0.2.

Whether the process actually proceeds depends on the storage of CO2, H2O, O2, and Glucose within the leaf.

During the night, the activity is reduced to 2 cores, representing the reduced metabolism.

Currently, the speed is set to a constant value, but there is the potential to adjust the speed or the number of cores to have variable productivity during the day.

Transpiration is further simulated as long as the stomata are open.

Currently, the tree’s available water amount is not limited, linking soil moisture values will be a crucial addition for future visualizations.

The final visualization illustrates the transfer of CO2, O2, and H2O from the tree to the atmosphere, which is determined by the measured VPD values over a single day-night cycle. The visualization highlight the impact of the water vapour treatment by showing a comparison between a control tree and another tree under treatment.