Within the UN-REDD Programme, the Food and Agriculture Organization of the United Nations (FAO) is working with various stakeholders and institutions in the Lower Mekong Region (LMR) to enhance forest governance, sustainability and legality of forest trade as well as data monitoring. In this context, FAO has worked with Viet Nam and other LMR countries to enhance capacities on an innovative forest monitoring system, making use of the recently added near-real-time monitoring capabilities on FAO’s SEPAL platform (sepal.io).

This article reflects on the experience of Viet Nam in testing this innovative tool by generating and verifying forest disturbance alerts from medium resolution satellite data.

Viet Nam example and lessons learnt

Illegal logging remains a pressing concern in Vietnam, even with the government's intensified enforcement measures. The trade in illegally sourced timber involves a diverse spectrum of actors, ranging from powerful, large-scale networks to small-scale operators, which affects forests throughout the country, including protected areas. In this context, Viet Nam, through a collaboration between FAO technical experts and the Forest Inventory and Planning Institute (FIPI) tested the feasibility and functionality of the SEPAL “change alert” tool across three provinces: Hoa Binh, Thua Thien Hue, and Gia Lai representing different forest ecosystems of the country. The selection of the Area of Interest (AOI) was based on both the diversity of forest types and the prevalence of illegal logging activities observed in the provinces over the past few years.

Location of area of Interest (AOI). Map Photo credit: FAO, Base map of Planet NICFI used in background along with global boundaries from QGIS

Disclaimer: The designations employed and the presentation of material in the map(s) do not imply the expression of any opinion whatsoever on the part of FAO concerning the legal or constitutional status of any country, territory or sea area, or concerning the delimitation of frontiers.

How Does It Work?

The SEPAL platform allows easy access to the vast amount of satellite data available through Google’s Earth Engine via an intuitive Graphical User Interface (GUI). Pre-defined workflows, called recipes, are at the heart of the platform and allow the users to quickly adapt state-of-the-art remote sensing techniques. At the same time the users can customize the parameter settings for local conditions, thus optimizing the results for their region of interest.

The change alert recipe is one of the latest developments and is targeted to support the users with generating forest disturbance alerts as quickly as possible. Those alerts can be generated on-the-fly by a combination of various satellite missions following a straightforward workflow. Apart from the medium resolution Landsat and Sentinel-2 missions, the tool is capable to deal with the daily high-resolution data from Planet. Also, cloud-penetrating radar data can be ingested and used for generating alerts independent of cloud coverage.

These alerts can then be further verified with (very) high-resolution images and, when possible, by teams checking the changes on the ground. The idea is that in an operational scenario verified alerts are then sent out to field teams that are responsible for forest monitoring and protection, safeguarding forest resources.

Image 1: Split view of the SEPAL change alert recipe, which allows for comparison of the use of different vegetation indices as well as (very) high-resolution data.

Image 2: example of alert generated using the “change alert” recipe, along with high resolution satellite images from before and after the event happened (blue circle). Photo credit: FIPI team, Viet Nam

Key findings

Adaptability is Key: Customization of main parameters such as the use of different vegetation indices led to varying results in each of the provinces. This underlines that there is no “one size fits all “approach and customization allow to improve the quality of the generated alerts, ultimately leading to a more efficient overall workflow

Accuracy Matters: The tool was highly accurate in Thua Thien Hue province, reliably identifying forest changes over 90% of the time. Results were less consistent in other provinces, highlighting room for improvement.

Forest Type Impacts Results: Evergreen forests and plantations were easiest to track. Although the tool struggled more with certain forest types like bamboo or seasonally changing forests.

Image 3: An example of how Viet Nam used the forest alert system tool. (a) Alerts as shown in the tool, (b) confirmation of the change in high-resolution imagery, and (c) verification on the ground. Photo credit: FIPI team, Viet Nam

Real-World Impact

The change alert recipe provides a way to turn the vast amount of satellite data into valuable and actionable information to support officers in the piloted provinces. Mr. Dang Quang Thuyen from FIPI, one of the 3 technical experts testing the tool, said that there are some remaining issues including the user-interface, but sees potential for the tool to be applied in the country. The ability to detect forest loss quickly translates to a faster response time, offering enhanced capabilities for forest protection.

Lessons for the Future

Viet Nam's experience demonstrates the potential of this near real-time monitoring tool while also showing areas for future development. As technology improves and teams gain more experience with these systems, we can anticipate even greater strides in forest protection efforts.

To learn more about other countries' experiences, find a previous article here.

Authors: 

Naila Yasmin 

Remote Sensing and GIS Expert 

FAO Forestry Division

Andreas Vollrath

Forestry Officer 

FAO Forestry Division

Erik Lindquist 

Senior Forestry Officer 

FAO Forestry Division

This article was originally published by UN-REDD Programme

Click for more