Big data is transforming forestry
Forest machines are expensive and having them stand idle is even more costly. It is difficult to predict the conditions of forest work, because they are affected by the weather, season, soil properties, elevation differences and the developmental phases of tree species. If forest work is done at the wrong time, it causes unfortunate environmental damage. Heavy machines can easily sink into clay soil.
The first step in preventing environmental damage is to understand the forest’s conditions and the changes that affect it.
Here, big data and new ways of creating prediction models based on big data, are helpful. Big data means the collection and analysis of extremely large amounts of data. The prediction models being developed in an EFFORTE project indicate when and with what kind of equipment it is worth going to the forest with.
“Prediction models are primarily a tool for planning forest work,” says Project Coordinator Jori Uusitalo.
Going into forests with fine-grained loam and silt soils, as well as peat soils, is done during the dry time in high summer or winter.
During the rest of the year, the work is done on ground with high bearing capacity.
“When the forest’s soil type and the properties of it can be assessed, work can be offered for forest machines all year round.”
Big data has already been collected previously on the ground’s elevation changes, for example. The EFFORTE project aims to combine different big data archives.
“For example, weather data combined with changes in elevation tells us where rainwater accumulates after a shower. Forest work can be planned more precisely with this information.”
Big data tells us how the forest is feeling right now. Soon, research groups and companies doing development will not have to separately collect their data. As measurement systems and sensors develop, the data will accrue while the machines work in the forest.
As measuring systems and sensors develop, systems will learn to predict conditions more precisely. The data that accumulates alongside forest work will improve and flexibly alter prediction models as conditions change.
Prediction models can be created on, for example, how ‘sinky’ the soil is or the quality of the forest’s trees.
The tracks left on the ground by a forest machine tyre are an indicator of how much the heavy machine sinks into the soil. The depth of the track can be measured with the help of the engine power.
”The more the tyre sinks, the more power the machine needs to move forward,” says Uusitalo.
The tracks, combined with weather measurements, let you know when it is a good time to head to the forest.
The quality of wood can be predicted by examining other trees that have been cut down in the area. The logger takes a photo of the felled tree’s stem. Counting the annual rings provides information on the growth rate of the tree. The harvester calculates the diameter and length of the trunk as the pruning and cutting proceeds. The result is called a trunk profile.
“The growth rate and trunk profile tell us about, for example, the wood’s density, firmness, moisture content and rottenness. The information helps us optimize what kind of products to make out of the wood.”
Smart ways of measuring and analysing big data will ultimately lead to the forester getting to know the forest better. Different conditions require different procedures.
“A forest hollow is humid, but the hillside is dry. A large forest is different in different corners,” Uusitalo explains.
Precise information on a forest and its soil helps in targeting the work. This is called precision forestry. The work can already be optimized when saplings are planted.
“Different tree species thrive under different conditions, so the forest’s soil, dryness, humidity and elevation differences have to be taken into account. If a species of tree thrives in dry surroundings, dry tussocks are built for its saplings in waterlogged parts of the forest,”, says Uusitalo.
Forestry is a chain containing many links, where each stage of work affects the costs of the next one. People may fail to notice certain details, which the prediction models are able to indicate in good time.
”When you understand the soil’s properties and growth potential, you are able to take into account the total costs and total return.”