How high-tech is training people and people are training high-tech
Figure 1: Photograph of detected site damage extent in Borneo, Indonesia.
A fire in the rainforest. Trees logged. Remnants burnt. Nothing happens. A field is prepared for agriculture. Wildlife flees. Still, nothing happens. What is going on? Doesn't anyone care? People do care and they do want to act. But rainforests are big and difficult to access. It can take over a year or more before a patrol manages to reach a certain area. What they need is information about what is happening, and where.
Space4Good is working closely with our partner Arsari Enviro Industri, based in East Kalimantan, Indonesia to monitor over 200, 000 hectares of land to better mitigate deforestation and fires. Our deforestation detection alert procedures, noted in our previous blog post here, combine our satellite data analysis algorithms with Fire Alerts provided by NASA, overlaid with GLAD and RADD deforestation alerts from Global Forest Watch. We then go on to validate incidents with our local implementation partners to best ensure that what we are detecting from afar is indeed happening up close.
In many cases noted to date in Indonesia, fire is not only fire. Very often we see fire as an indicator of much more. Deforestation, clear-cutting, encroachment, and in most cases illegal activities. It is in the combined efforts of Space4Good deforestation and fire detections and our partner mitigation practices that we begin to discover the powerful insights that can be unlocked by remote sensing technologies.
As soon as we get a fire or deforestation alert that is inside the monitored area or closeby, an alert is generated and disseminated to in-field operatives via a navigation application on the phone. This is also able to be used offline to ensure nothing is missed. Additionally, messages are pinged to WhatsApp groups to alert fieldworkers. It is here that the real work begins! Field operatives then navigate their way to the identified locations and validate if the event occurred (or not). Our partner will then inspect the area and send us their feedback with the location of the event, the description of the occurrence, and photos taken by phone (and in some cases, drone to ascertain a better understanding of the extent of the incident). These incident validation reports then provide a more complete overview of the kinds of activities and extent of the damage.
The following graph shows a summary of the steps taken to best inform our partner networks when an event is detected:
Figure 2: Summary of the current procedure followed for fire and deforestations detections.
Thanks to this cooperative dynamic, we can keep track of the events and improve our algorithms to go on to identify various trends. In identifying these trends, we unearth a wealth of information otherwise not taken into consideration and or previously unknown. This perspective has provided us and our partners the opportunities to learn interesting correlations.
For example, as we began to map and ascertain drivers for these activities, we noted that road accessibility and the proximity to villages play a huge role in the locations of fires. As accessibility increases to forested areas, the possibility of swift and extensive timber extraction becomes easier and in this case, fires are a sure-fire giveaway. In other cases, encroachment has been detected alongside roads where palm oil trees were planted illegally. Remote sensing was able to provide the perspective necessary to see these changes, and also determine that this had been happening for quite a while. Without remote sensing, this site could have gone undetected for much longer.
When considering these events in line with historical data, we are also able to see correlations between deforestation with a lack of security in the area. As a result, by assessing the events over a period of time, we are able to identify the vulnerable areas and advise on where to focus resources to better mitigate these crimes.
Figure 3: Example of validation images confirming detected deforestation and fire event in Borneo, Indonesia, 2020.
Conversely, even when teams are deployed to safeguard these forests, corruption finds its way in. Very often we even see intra-organizational breaches, meaning people working within the local organizations designed to protect the area are those who then provide illegal loggers and land encroachers access. By correlating staff locations with activities over time, pattern recognition and cross-referencing can better ensure that these breaches can be stopped, are less likely to happen, and simultaneously increase transparency within our partner organizational networks.
This social impact seen in monitoring this region does not stop there. Based on feedback from field operatives, equipment and motorcycles from the perpetrators have been discovered abandoned in the area. This indicates the activities of the wood thieves recruited by syndicates - usually coerced through economic exploitation or gifts of vehicles later on abandoned due to lack of money to maintain them. This exemplifies the degree of investment these criminal networks will engage in to reach these rare and valuable woods. A sad and very real truth.
In the end, monitoring the area and having these insights is only one piece of the story. It is in partner relationships that necessary ground-truthing for accurate algorithm development can be made, implementation and mitigation procedures can be carried out and ultimately, the real impact can be ensured. Since starting in 2019 our joint approach has successfully detected 573 deforestation detections (Since 2019) and 112 fire detections (Since 2020) in the 200,000 hectares Mixed Tropical Forests monitored.
Would you like more information or are you interested in collaborating with Space4Good? Visit our website or contact us via firstname.lastname@example.org we are happy to assist.