This R repository receives the output of the TMF implementation (https://github.com/quantifyearth/tmf-implementation) for N REDD+ projects to be analysed, and calculates the baseline deforestation and carbon loss, generates additionality forecast and evaluates overclaiming risk. The baseline is currently based on the set M (matches.parquet), and is subject to change in the future.

The core script, forecast.R, contains the following five parts: 0. Setup A. Obtain observed additionality B. Predict deforestation probability of baseline pixels using logistic regression C. Calculate boostrapped baseline C loss D. Generate additionality forecast and estimate overclaiming risk

It requires the following input variables to read TMF implementation output and other data. All variables are vectors containing one value for each project to be analysed:

1. projects: an index of all projects to be analysed This should correspond to the filenames of the shapefiles and to the -p argument in the implementation code It is usually be the ongoing projects' VCS ID or customised (e.g. prefixed series of integers)

2. pair_dirs: absolute paths of the directories containing all matched pair pixel sets (typically "/pairs/xxx.parquet" and "/pairs/xxx_matchless.parquet") The directory should containing pairs of parquet files with the same file name, with and without the "_matchless" suffix. This is used to calculate estimated observed additionality.

3. k_paths: absolute paths of the set K (typically "k.parquet")

4. m_paths: absolute paths of the set M (typically "matches.parquet") Both should be in parquet format, containing the following columns: "lat", "lng" (degrees), "slope" (degrees), "elevation" (metres), "access" (remoteness, in minutes), "cpc10_u", "cpc10_d", "cpc0_u", "cpc0_d" (from 0 to 1), "luc_[t-10]" to "luc_2021" (categorical, 1-6, based on the JRC-TMF dataset, Vancutsem et al. 2021), "ecoregion" (categorical, based on the RESOLVE dataset, Dinerstein et al. 2017)

5. acd_paths: absolute paths of the carbon density per LUC (typically "carbon-density.csv") This should be an csv file (although the script could be modiified in the future to support txt format) containing columns "land.use.class" (categorical, 1-6) and "carbon.density" (MgC/ha) for all six LUCs, although the script checks and fill missing LUC with NAs

6. polygon_paths: absolute paths of the shapefile of the project extent This should be a geojson file containing valid geometries in WGS84 (EPSG: 4326), although the script checks for both conditions. This is currently only used to calculate project area (ha), but could be useful for other purposes in the future.

7. country: country of the project

8. t0: year of start of the project (real or hypothetical)

9. OPTIONAL: proj_ID: ID used to identify country t0 from the proj_info_all data frame This is only used to remove the trailing "a" that I added to the filename of some shapefiles that needed fixing If unspecified, the projects variable will be used #10. out_path: absolute paths of the directory where outputs are to be saved; include file prefix if desired

It generates the following output:

A.

1. project_var: project-level variables This is a csv file containings N rows (one for each project), and the columns "project", "t0", "country", "area_ha", "acd_undisturbed", and the min/max/median over 100 samples of the following variables: "slope", "elevation", "accessibility", "cpc{0, 5, 10}_ {u, d}" (coarse proportional cover at t0/t-5/t-10 of undisturbed/deforested pixels), "defor_ {5_0, 10_0, 10_5}" (deforestation defined as CPC change over the period of t-5 to t0, t-10 to t0, or t-10 to t-5).

2. OPTIONAL output: only basic project-level variables "project", "t0", "country", "area_ha"

3. additonality_estimates: observed additionality time series It is currently a list of N element, saved as an RDS file that can be read as an R object. Each element is a data frame containing the columns "year", "c_loss" (counterfactual carbon loss), "t_loss" (project carbon loss), "additionality", "pair" (index of the sampled pairs), "started" (FALSE for years before and including t0, TRUE for years after t0). Each row is a combination of a year and a sampled pair.

B.

1. baseline: predicted baseline deforestation probability It is currently a list of N elements, saved as an RDS file that can be read as an R object. Each element is a data frame containing the columns "lat", "lng", "slope", "elevation", "access", "luc10", "luc0", "cpc10_u", "cpc10_d", "cpc0_u", "cpc0_d", "defor" (TRUE if land use class has changed from 1 to 2, 3, or 4 between t-10 and t0, FALSE otherwise), "defor_prob" (predicted deforestation probability based on logistic regression, between 1 an 0), "risk" ("low" if defor_prob < 0.01, "high" otherwise), "acd10", "acd0", "c_loss" (annual carbon loss rate). Each row is a baseline pixel.

2. #project_defor_prob: predicted project deforestation probability It is currently a list of N elements, saved as an RDS file that can be read as an R object. Each element is a data frame containing the columns "lat", "lng", "slope", "elevation", "access", "luc10", "luc0", "cpc10_u", "cpc10_d", "cpc0_u", "cpc0_d", "defor" (TRUE if land use class has changed from 1 to 2, 3, or 4 between t-10 and t0, FALSE otherwise), "defor_prob" (predicted deforestation probability based on logistic regression, between 1 an 0), "risk" ("low" if defor_prob < 0.01, "high" otherwise), "acd10", "acd0", "c_loss" (annual carbon loss rate). Each row is a project pixel.

3. range_defor_prob: total range of predicted baseline deforestation probability in baseline pixels It is a data frame of N columns, each for a project. Each column contains the minimum and maximum of predicted baseline deforestation probability.

4. baseline_summary: basic information about the baseline This is a csv file containing N rows (one for each project), and the columns "project", "baseline_area" (in number of pixels), "low_risk_ratio" (ratio of baseline pixels classified as "low-risk" based on a threshold of 1% predicted deforestation probability), and "slope", "elevation", "access" (indicating what is the logistic regression result for each of the environmental variables: "Neg." for negative effect, "Pos." for positive effect, and "N.S." for non-significant effect)

C.

1. df_ses: standardised effect size of change in bootstrapped carbon loss by using only high-risk pixels instead of all pixels in This is a csv file containing N rows (one for each project), and the columns "project", "ses" (standardised effect size). The standardised effect size is calculated as ("mean carbon loss rate in high-risk pixels" - "mean carbon loss rate in all pixels") / "standard deviation of carbon loss rate in all pixels"

2. c_loss_boot: bootstrapped baseline annual carbon loss rates for all projects This is a csv file containing a long-form data frame, containing the columns "project", "type" ("all" for all pixels, "high_risk" for only high-risk pixels), and "val" (each bootstrapped carbon loss rate). The number of values generated (number of rows in each project-type combination) is defined by the variable "boot_n"

D.

1. forecast_summ: additionality forecast under different scenarios This is a csv file containing N rows (one for each project), and the columns "mean_100", "ci_100", "mean_75", "ci_75", "mean_50", "ci_50", "mean_25", "ci_25", and "project". The columns prefixed "mean_" indicate the forecasted mean annual additionality (MgC/ha/yr), and the columns prefixed "ci_" indicate its confidence interval, under the scenarios of 100%, 75%, 50%, and 25% project effectiveness, respectively.

Apart from these outputs, the script also generates a number of plots for visualisation. Currently they are all used for figures in E-Ping's manuscript, and can be deactivated by setting the variable "visualise" to FALSE.