Forestry and Forest Products Research Institute.
REDD Research and Development Center.

Text Size

Development of measurement method for emission reductions

Companies and other entities need to reduce the costs required for monitoring emission reductions when implementing REDD+ activities under the JCM. However, accurate estimation of emission reductions is necessary to secure the credibility of credit. In addition, monitoring in cooperation with local communities and local governments is indispensable to implement REDD+ activities smoothly.

Therefore, we present how to select the optimum combination from the methods of measuring emission reductions suitable for the project level under the JCM according to local conditions such as national requirements and forest ecosystem, social environment, and the condition in the host country. At the same time, we demonstrate the procedure to apply and implement the selected measurement method to estimate the reference level and how to transfer technology to local communities and local government and how to establish cooperation. For this purpose, we studied the classification of forests necessary for monitoring at the local level and the influence of the number of ground truth data necessary for classification and verification on the estimation of forest area change required for estimating emission reduction and examined the survey method to improve ground-based survey. With this result, we have developed an optimal method to measure emission reductions in forests under various environmental conditions. In order to address these challenges, we selected Myanmar, Cambodia, and Peru as target countries, taking account of their interest in REDD+, reduction potential, possibility under JCM, difference in forest ecosystem and social environment, and ripple effect on neighboring countries.

Here, we developed technology and implemented a case study under the following subtheme.

  1. Establishment of measurement (monitoring) method of emission reductions in forests under various environmental conditions
  2. Establishment of measurement (monitoring) method of emission reductions in forests under various environmental conditions
  3. Establishment of verification system for measuring (monitoring) results of practical emission reductions

1 Establishment of measurement (monitoring) method of emission reductions in forests under various environmental conditions

(1) Introduction

When measuring emission reductions through REDD+ activities under the JCM, that is when monitoring changes in forest carbon stocks, we consider that the method combined remote sensing and ground-based survey is effective as UNFCCC recommends to implement monitoring at the national level. In forest observation using remote sensing, different image processing and analysis methods are required depending on the natural environment such as ecoregion, seasonality, and topography, land use, institution and other social environments. In this subtheme, we have develop methods to measure emissions reductions targeting from the Andes ecoregion in Cusco to the highland Amazon ecoregion where vegetation types and tree stands are different depending on the altitude and influence by topography in cooperation with Ministerio de Agricultura y Riego ? Servicio Nacional Forestal y de Fauna Silvestre (MINAGRI-SERFOR). Table 1 shows the development plan of the measurement method and concrete activities. In addition, we conducted a ground truth for analyzing satellite data at 131 points, and studied how to determine biomass for estimating biomass and its change using satellite image. In FY2016, in order to study the biomass mapping method for extracting biomass change in the target site, we conducted a plot survey at 26 places and clarified the species composition in the degraded forest.

Table 1 Plan for developing measurement method (3 years) and concrete activities

2015 2016 2017
Understanding of distribution of biomass and land use in target area Consideration on biomass mapping method for extracting biomass variation Establishment of measurement method considering altitude, topography, human-induced disturbance
・Plot survey at 20 points
・Plot survey at 20 points
・Plot survey at 20 points
・Plot survey at 100 points
・Plot survey at 100 points
Plot survey at 100 points
・Preliminary analysis on biomass and species composition
・Elucidation of changes in species composition in degraded forest
・Elucidation of the relationship among biomass and tree stands and species composition
・Preliminary classification on the land use of the target area using satellite images
・Study on the method of setting biomass class for estimating biomass and its change through satellite image
・Organization of biomass mapping methods in different conditions

(2) Implementation system of research

We concluded MOU with SERFOR on Sep. 11, 2015 to address this challenge in the Republic of Peru. Under this MOU, we appointed project coordinators, remote sensing coordinator, and coordinator for ground biomass measurement in both institutions for smooth implementation of the project. In this year, the project coordinator has been changed due to the change of administration in Peru.

  • Collaboration Coordinator:
    Rocio Malleux Hermani (SERFOR)
    Yasumasa Hirata (FFPRI)
  • Coordinator of Remote Sensing Studies:
    Lenin Cruyff Ventura Santos(SERFOR)
    Vega Isuhuaylas Luis Alberto (FFPEI)
  • Coordinator of Ground-based Inventory Studies:
    Alexs Arana Olivos(SERFOR)
    Tamotsu Sato (FFPRI)

Related local office:Regional office of Cusco department, SERFOR

(3) Clarification of species composition in degraded forests

The REDD Research and Development Center has conducted a survey on biomass in various forest types in Asia and South America to gather knowledge related to the methodology of the ground-based survey for implementing REDD+. For example, we have developed a ground-based survey method for estimating biomass of tropical rain forests and tropical seasonal forests through the surveys in the Malay Peninsula and the Indochinese Peninsula (Thailand/Cambodia). In South America, we conducted a survey for estimating biomass targeting open forests in a very dry savanna climate in Paraguay. In South American forests, as the altitude increase, the tropical forests, as typified by the Amazon, change to the forests of the Andes mountain range. In the Andes mountain range where mountains exceeding 6,000 m in altitude are connected, forest structure is diverse and mountains from the low altitude zone (about 1,000 m in altitude) connected to the tropical forests to the high altitude (3,500 m in altitude) where a very few tree species grow, are continuously distributed. Furthermore, the forests have been affected by human activities since ancient times; therefore, many degraded forests are seen. In this project, we aim to establish a method for estimating forest carbon stocks in mountainous areas including that in high altitudes based on the experience in tropical forests and dry forest to date. In this fiscal year, we established plots for ground-based survey and carried out a complete enumeration, and tried to figure out the biomass distribution and identify the factors that affect the species composition in forest stands.

We targeted the forests in the Cusco province of the Republic of Peru, where the altitude is approximately 700-3,500 m. Ecoregions in this area are largely classified into "upland tropical forests (Selva Alta)" and "montane forests (Sierra)". The boundary between these two ecoregions is not clear and varies depending on the basin, but generally divided at approximately 1,000 m. We established a survey plot (a quadrat of 40 m × 40 m) in the target area for conducting a complete enumeration mainly for live trees with diameter at breast height of 10 cm or more in the plot. Then, recorded the condition of the forest (topography and slope), the history of natural disturbances (erosion, etc.) and human-induced disturbances (fire and cultivation of crops, etc.) as much as possible. We also categorized the forest type as "disturbed primary forest" (hereinafter referred to as primary forest), "secondary forest" and "plantation forest". In order to reduce variation in measurement accuracy, every time a new staff member participate in SERFOR, we confirmed the survey method using a simple manual (Spanish version). In addition to 20 forest stands surveyed in FY2010 (13 forest stands for primary forest and 7 forest stands for secondary forest), we examined 26 forest stands (5 forest stands of primary forest, 20 forest stands of the secondary forest, 1 forest stand of the broadleaved plantation forest).

In order to calculate carbon stocks per unit area, we estimated biomass in the forest stand based on the data set of the above 46 forest stands. In the survey results of the last year, the above-ground biomass was clearly different between the primary forest and the secondary forest, and the biomass of the primary forest was greater, but in the secondary forest, the relationship between the altitude and the above-ground biomass that was observed in the primary forest was not clearly detected. In this survey, it was found that in the secondary forest the biomass gradually decreases as the altitude increases. However, the decreasing trend was different between the two types of forests. In the primary forest, the above-ground biomass decreased sharply as the altitude increases as compared with the secondary forest.

In the survey conducted in FY 2015, identifying individual tree species was difficult due to the lack of experts in the host country. This situation was improved greatly in FY2008 as a new SERFOR staff member who has high ability to identify tree species took part in this survey. The species, which have known only by the local name, were recorded by scientific names. Also, all species that have been recorded by local names were recorded by scientific names and tress species that could not identified even their family names have drastically decreased. However, tree species that can identify up to the specific epithet is limited, and in many cases, they were identified up to the genus level.

The community composition of the forest stands was different between the primary forest and the secondary forest (Fig. 1). In both forest types, the most dominant tree species was the legume family (Fabaceae). For other dominant species, Lauraceae was found in the primary forest, and pioneer tree species such as Urticaceae, Cannabaceae and Euphorbiaceae were found in the secondary forest. The dominant tree species in the second forest are low in wood density, thus, the above-ground biomass is lower than the primary forest. In this survey, in the plantation forest surveyed in one forest stand, agroforestry was performed, under a canopy tree, Inga adenophylla, which belongs legume family, the Avocado (Persea Americana, -Lauraceae) and coffee trees were cultivated.

We ranked the forest stands based on species composition (genus level) using Nonmetric Multi-Dimensional Scaling (NMDS). The arrangement of forest stands developed on the two axes can mainly be explained by difference in altitude and above-ground biomass. However, the factors that explain the axis includes not only natural disturbance but also human-induced disturbance. This can be said that it is useful to record information on the disturbance history of forest stands during the complete enumeration as much as possible. In order to confirm whether the result of this survey has a certain degree of generality, it is necessary to collect data in the future.

Fig. 1 Comparison of tree species composition between forest types (5 families of the highest relative abundance based on basal area)

(4) Consideration biomass mapping methods for extracting biomass variation

In order to examine the biomass mapping method for extracting the change in the amount of biomass, we carried out a ground truth at 131 sites from late June to October 2016 and analyzed it along with the survey data at 160 sites conducted in FY 2015. We planned the survey at SERFOR's headquarters in Lima and arranged cars and workers and checked the road conditions in the local office in Cusco. The Andes ecoregion in Cusco and the highland Amazon ecoregion were selected for the survey. In the survey sites, as conducted in 2015, we confirmed the land cover and conducted survey by counting trees applying the Bitterlich method in order to estimate the carbon stocks in forests (Fig. 2). Among 70 survey sites in the forests, 53 sites are natural forest and 17 sites are plantation forests.

We recorded the survey results in the sheet prepared for the ground-truth. After the survey, we converted it to an electronic file with a scanner to save it and to share with both institutions, and also recorded it to an excel sheet to create a data base.

  • Preparing for survey

  • Confirming survey procedure

  • Ground-based survey by
    counting trees

  • Measuring the counted trees

Fig. 2 Ground truth