2.1 Three Excel-based forest ecosystem services (ES measuring and monitoring tools)

Three Excel-based tools were developed under the USAID LESTARI project (2015–2020) by Michigan State University: a Forest Carbon (FC) tool, a Tree Biodiversity (TB) tool and a Forest Integrity Assessment (FIA+) tool that includes functions for assessing forest resources used for cultural and provisioning benefits by local people. The tools store and manage field-based measurement and observation data. The tools also automatically compute data based on the field data inputs.

2.2 Phased training program for KPH staff and Social Forestry community members

A three-phase training program was implemented to build the capacity of KPH staff and Social Forestry community members. Phase 1 training focused on: introducing participants to the fundamental concepts of ecosystems and ecosystem services, review of several scientific phenomena (carbon cycle, greenhouse effect, climate change), explanation of the five pools of measurable forest carbon, the use of allometric equations for measuring biomass, field measurements and sample plot designs for forest inventory, and a preview of the three ES tools.

Phase 2 was a field-based practicum where participants were given hands-on training in demarking fixed-area nested plots and collecting plot-level data for use in the three tools. Data collected included: tree dbh, measurements to compute tree height (distance from tree, angle of clinometer, height above ground to observer’s eye level), tree species counts, seedling counts in 2×2 m plots, and FIA+ checklist observations.

After participant teams collected field inventory data, a Phase 3 training was convened and focused on understanding the results of the Tool analyses and how best to use each of the ES Tool’s information for developing improved forest management plans.

2.3 Field data collection protocol

KPH and Social Forestry community teams collected field data for use in each of the three ES tools using a systematic protocol for plot sample locations, plot design and data collection procedures. Data collection teams ranged from 3 to 10 members.

Field plot locations were established on a regular grid pattern. Each plot was located at a minimum distance of 100 m from any forest edge and 50 m from rivers or streams. The plots were located at a minimum distance of 50m from each other. No plots were located in non-tree areas.

Field plots for the FC tool were three square nested plots in which specific diameter class trees were measured: 20 x 20 m plot for trees greater than 20 cm dbh, 10 x 10 m plot for trees less than 20 cm and greater than or equal to 10 cm dbh, and 5 x 5 m plot for trees less than 10 cm and greater than or equal to 5 cm dbh. Seedling counts by species were recorded from two 2 x 2 m plots located outside the 20 x 20 m plot. Within the three nested plots, tree species, dbh, and data for calculating tree height were recorded for all trees ≥ 5 cm dbh.

The same plot was used to also record data for use in the TB and FIA+ tools. Total tree species counts for all trees ≥ 5 cm dbh within the 20 x 20 m plot were recorded for the TB tool. Observations of forest structure, impacts, focal habitats, focal species as well as forest resources noted as used by local people for provisioning and cultural uses were recorded while standing inside the 20 x 20 m plot. All noted observances of these indicators and species within the plot or seen outside the plot while observing from inside the plot were noted.

Field data were recorded on paper field sheets and then input to the appropriate ES tool following data collection.

3.1 Forest carbon and tree biodiversity in nine forest areas in Aceh Province

We summarize the forest carbon average (Mg C ha^-1) and total carbon stock (Mg C) listing the allometric equation used and the tree biodiversity species richness, evenness and dominant species for the nine forest areas in Table 4. The results from the FC and TB tools are what we would expect. Higher average carbon per hectare values in the tropical evergreen and mangrove forest types than in the coffee agroforestry or pine forest types. The range is from a high of 446.93 Mg C ha^-1 in the tropical evergreen social forestry of HKm Rumah Rungkoh to a low of only 27.14 Mg C ha^-1 in the young coffee agroforestry area. Also, low biodiversity in the mangrove and pine forests than what is measured in the tropical evergreen forests. HKm Rumah Rungkoh and the two urban forest areas, Kota Langsa and Kota Nagan Raya, are considered secondary tropical evergreen forest (35, 33 and 33 species richness, respectively, with evenness values close to 1: 0.86, 0.79 and 0.84). The evenness values close to 1 indicate that the forest is not dominated by just one or two species. The two mangrove areas and the pine forest area have very low species richness values, 1, 2 and 3 species only, and their evenness values are closer to zero, with the pine forest evenness the lowest at 0.06.

The overall forest carbon stock and tree diversity are important in general for a forest area, but perhaps equally important is the spatial variation of the forest carbon and tree diversity which is information that can be useful for forest resource management activities. Understanding the variation of carbon and tree biodiversity across the plots can assist in developing management plans and actions not just in general for the forest area as a whole but to specific areas within the forest itself.

Table 5 shows the range of carbon and species diversity for the nine forest areas. Again, as expected, the mangrove forest areas and also the pine forest areas have low species richness values and, therefore, the maximum-minimum range between plots is small (mangrove 1–3 species and pine 1 species). Noteworthy from these plot level species richness data in comparison to species richness of the forest areas as a whole are the tropical evergreen forests. While HKm Rumah Rungkoh and the two urban forests have species richness values greater than 30, none of the plots in any of these forest areas include 20 or more species. Some of the tree species, therefore, are present only in some of the plots. Furthermore, in the two urban forest areas, there is at least one plot in each forest area with only three species present and in HKm Rumah Rungkoh, only eleven species present. These forest areas, while overall having a high evenness value, are not uniform in their diversity across the landscape. This is true also for forest carbon. The range in plot-level carbon is greatest in one of the mangrove forest areas with a maximum-minimum value of 667.00 Mg C ha^-1. The forest area with the smallest range is the coffee agroforestry area, with a maximum-minimum range of only 36.78 Mg C ha^-1 between plots.

3.2 Forest integrity and health in nine forest areas in Aceh Province

The FIA+ tool provides a more qualitative assessment of the health and integrity of the forest ecosystem scoring a forest by the presence or absence of specific attributes. These include attributes such as the presence of large trees or the absence of stumps which indicate impacts from logging. Teams from seven of the nine forest areas were able to collect field data for the FIA+ tool. The presence or absence of focal species in the forest area is also an indicator of a forest ecosystem’s health. It is presumed that in highly degraded, poor forest ecosystems, there are few focal species. Since many faunae focal species can be difficult to observe directly, data collectors record the presence of a focal species through other means as well, such as a species nest, their call or song in the case of a bird, their marks as in the case of a sun bear, or their faeces. Table 6 includes the average, minimum and maximum FIA+scores at the plot level, the number of focal species observed and the number of observations for the focal species list.

The mangrove forest in Telaga Tujah has the highest average FIA+ score, followed closely by HKm Rumah Rungkoh, 32.7 and 30.8, respectively, out of a total of 40 points possible. Desa Agusen LPHD Gembuluh Berkah, a coffee agroforestry area, had the lowest score of 8.2, indicating a highly degraded forest, which is supported by the low number of focal species and observations of focal species, four and nine, respectively. HKm Rumah Rungkoh has 11 focal species present with 42 observations and the mangrove forest in Telaga Tujah has seven species with 62 observations. Hutan Kota Langsa also has 11 focal species present with 27 observations; however, this urban forest is also a zoo. Table 7 lists the focal species observed for each forest area.

3.3 Carbon stock estimate accuracy and error

The limited number of plots in some of the forest areas and the wide variation in carbon values among these plots do impact the statistical accuracy for reporting carbon stock in the full forest area. The FC tool includes several metrics to understand the accuracy of the forest carbon stock estimates. Both the standard error and sample error are reported in the FC tool for the field plot inventory data. The FC tool also reports the lower and upper bound average Mg C ha^-1 at the 95% confidence interval. These values are used to compute a lower and upper total carbon (Mg C) stock value along with the estimate computed from the average Mg C ha^-1 for all plots. The tool also reports, based on the plot data input (given the standard and sampling errors, the mean, and standard deviation of the plot carbon (Mg C ha^-1)), an estimated total number of plots to meet an accuracy of 95% confidence interval (CI) at 10%, 15%, and 20% error. Table 8 shows the two error values, mean plot carbon, upper and lower bound plot carbon, and the number of field plots together with the estimated number needed to meet an accuracy of 95% CI at 10%, 15% and 20% error for each of the nine forest areas.

The range of carbon values at the plot level is an obviously important factor. For any set of plots where the standard deviation is close to the mean carbon value, this, too, is an important factor. Only two forest area carbon estimates have a sufficient number of plots with low enough sampling and standard errors for accuracy at 95% CI and better than 15% error. These are Desa Agusen LPHD Gembuluh berkah and Pinus Group Reje Pudung/Terangun Gayo Lues. Two areas have an accuracy right at 95% CI and 20% error: HKm. Rumah Rungkoh and Mangrove Telaga Tujuh. The rest of the forest areas estimate carbon stock values with an accuracy of greater than 20% error at 95% CI. In the case of the mangrove forest with only four plots, there is such high variation in the plot carbon values that another 144 plots would be needed for an accuracy at 20% error, 95% CI (assuming no change to the mean carbon and standard deviation with the inclusion of the additional plots).

In the forest areas with fewer plots than required for significant accuracy, the lower bound estimate of carbon stock should be reported. It may also be prudent to stratify the forest area to help lower the variation of carbon at the plot level when estimating the forest carbon.