diff --git a/docs/source/workflows/drivers.rst b/docs/source/workflows/drivers.rst index 30e9c7036e..1edc779dd4 100644 --- a/docs/source/workflows/drivers.rst +++ b/docs/source/workflows/drivers.rst @@ -3,22 +3,24 @@ Direct drivers assessment .. note:: - This documentation has been produced in the framework of the global project, Assessment of deforestation and forest degradation and related direct drivers using SEPAL, supported by the Central African Forest Initiative `(CAFI) `__. + This article was produced in the framework of the global project, Assessment of deforestation and forest degradation and related direct drivers using SEPAL, supported by the `Central African Forest Initiative (CAFI) `__. - The Deforestation and Degradation Drivers (DDD) project is a global methodology developed and piloted in central Africa. While the Congo Basin is used as an example, this methodology can be applied to any geographic area. + The Deforestation and Degradation Drivers (DDD) project is a global methodology developed and piloted in Central Africa. While the Congo Basin is used as an example, the methodology can be applied to any geographic area. Background ---------- -Deforestation and forest degradation are complex processes that have many underlying causes. A comprehensive understanding of forest conversion to other land uses is instrumental for the development of policies and actions aiming to reduce the loss of forests and associated carbon emissions. It is important to understand recurring patterns and correlations that can help countries tailor their efforts toward reducing forest loss. The causes of deforestation and forest degradation vary both regionally and temporally.\ :footcite:`curtis:2018` Different studies refer to agricultural expansion (cropland and pasture) as the largest direct cause of global deforestation.\ :footcite:`fra:2022,gibbs:2010,hosonuma:2012,kissinger:2012,sandker:2017` Commercial agriculture is estimated to be responsible for around 70 percent to 90 percent of worldwide deforestation. In Africa, both commercial and subsistence agriculture may account for similar importance in terms of forest loss, while fuelwood collection, charcoal production, and to a lesser extent, livestock grazing in forests, are the most important drivers of degradation. +Deforestation and forest degradation are complex processes that have many underlying causes. A comprehensive understanding of forest conversion to other land uses is instrumental for the development of policies and actions aiming to reduce the loss of forests and associated carbon emissions. It is important to understand recurring patterns and correlations that can help countries tailor their efforts toward reducing forest loss. -However, these studies (as well as existing national studies on the drivers of deforestation and forest degradation in central Africa) are generally based on data acquired up to 2014 and do not consider the recent trends in observed tree cover loss. They also do not take into account the importance of the spatial fragmentation of forests and the role played by degradation induced by forest exploitation.:footcite:`molinario:2015` Furthermore, they simplify causes of deforestation into a single driver, which does not adequately reflect the complex local context and interacting causes – decisions that lead to anthropogenic forest disturbance at local levels.:footcite:`ferrer-velasco:2020` +The causes of deforestation and forest degradation vary both regionally and temporally (Curtis, 2018). Different studies refer to agricultural expansion (cropland and pasture) as the largest direct cause of global deforestation (FAO, 2022; Gibbs, 2010; Hosonuma, 2012; Kissinger, 2012; Sandker, 2017). Commercial agriculture is estimated to be responsible for around 70–90 percent of worldwide deforestation. In Africa, both commercial and subsistence agriculture may account for similar importance in terms of forest loss, while fuelwood collection, charcoal production, and livestock grazing in forests (to a lesser extent), are the most important drivers of degradation. + +However, these studies (as well as existing national studies on the drivers of deforestation and forest degradation in Central Africa) are generally based on data acquired up to 2014 and do not consider the recent trends in observed tree cover loss. They also do not take into account the importance of the spatial fragmentation of forests and the role played by degradation induced by forest exploitation (Molinario, 2015). Furthermore, they simplify causes of deforestation into a single driver, which does not adequately reflect the complex local context and interacting causes – decisions that lead to anthropogenic forest disturbance at local levels (Ferrer-velasco, 2020). These recent trends and the lack of updated studies result in little consensus on the main drivers and agents of forest change at regional levels. There is a pressing need to systematically quantify the direct causes of deforestation and degradation via a systematic and comprehensive approach. An updated assessment should provide validated evidence on the role and weight of different drivers of forest loss and support decision-making to address related challenges. A spatially explicit approach should also facilitate the assessment of the efficiency of policies and actions in different contexts. Improved spatial data on deforestation and forest degradation, as well as an improved understanding of the drivers, will support land-use planning approaches in two pilot areas where the regional analysis indicated a clear opportunity for supporting land-use planning and decision-making. In this context, the Food and Agriculture Organization of the United Nations (FAO) has developed a robust methodology to assess the recent trends and drivers of deforestation and forest degradation. It involves collaborative approaches in which national experts, research institutes and civil society work together and compile resources and data to reach a common view on the current and future causes and drivers of anthropogenic forest disturbances. These tools generate improved information for decision-making and land-use planning in relevant management contexts. -FAO, in partnership with CAFI, as well as a coalition of donor and partner countries, have developed a standard, global, large-scale methodology to assess forest dynamics using cloud-computing solutions and open-source tools. The approach maps disturbances (both deforestation and degradation) and quantifies the contribution of multiple associated direct drivers. The methodology is used to assess deforestation and forest degradation trends, and their associated current and historical direct drivers in six central African countries as part of an international effort to mitigate climate change, support the Sustainable Development Goals (SDGs), and protect biodiversity. The project builds on a collaborative approach, in which national experts, global research institutes, and civil society work together, as well as compile resources and data, to provide technical evidence and reach a common view on the current and historical trends, as well as direct drivers of forest disturbances. +FAO, in partnership with CAFI, as well as a coalition of donor and partner countries, have developed a standard, global, large-scale methodology to assess forest dynamics using cloud-computing solutions and open-source tools. The approach maps disturbances (both deforestation and degradation) and quantifies the contribution of multiple associated direct drivers. The methodology is used to assess deforestation and forest degradation trends, and their associated current and historical direct drivers in six Central African countries as part of an international effort to mitigate climate change, support the Sustainable Development Goals (SDGs), and protect biodiversity. The project builds on a collaborative approach, in which national experts, global research institutes, and civil society work together, as well as compile resources and data, to provide technical evidence and reach a common view on the current and historical trends, as well as direct drivers of forest disturbances. Definitions ----------- @@ -38,7 +40,7 @@ Forest definitions There are a total of four different national forest definitions in the six countries of the study region. These are applied at country level using information on canopy height, tree cover and pixel area. .. csv-table:: - :header: Scale, Source, Date, Minimum mapping unit (MMU) (ha), Tree height (m), Canopy cover (%), Comment + :header: Scale, Source, Date, Minimum mapping unit (ha), Tree height (m), Canopy cover (%), Comment Cameroon, "REDD+ National Strategy (MINEP, 2008)", 2018, 0.5, 3, 10%, "Exclusion of monospecific agro-industrial plantations with a purely economic vocation and using mainly agricultural management techniques; are always considered as forest the areas subjected to natural disturbances which are likely to recover their past status." Central African Republic, FRA, 2020, 0.5, 5, 10% @@ -62,8 +64,8 @@ The baseline map for the regional forest cover was first derived from a common c 2, Forest, Dense dry forest, Forêt Dense Sèche, Bosque denso seco, "Dense dry forest, >60% tree cover, with dry seasons" 3, Forest, Secondary forest, Forêt Secondaire, Bosque secundario, "Open forest, 30–60% tree cover, degraded or secondary" 4, Forest, Dry open forest, Forêt Claire Sèche, Bosque claro Seco, "Dry open forest, 30–60% tree cover, with dry seasons" - 5, Forest, Sub-montane forest, Forêt Sub-Montagnarde, Bosque sub-montañoso, "Forest >30% tree cover, 1100-1750m altitude" - 6, Forest, Montane forest, Forêt Montagnarde, Bosque montañoso, "Forest >30% tree cover >1750m altitude" + 5, Forest, Sub-montane forest, Forêt Sub-Montagnarde, Bosque sub-montañoso, "Forest >30% tree cover, 1100-1750 m altitude" + 6, Forest, Montane forest, Forêt Montagnarde, Bosque montañoso, "Forest >30% tree cover >1750 m altitude" 7, Forest, Mangrove, Mangrove, Manglar, "Forest >30% tree cover on saline waterlogged soils" 8, Forest, Swamp forest, Forêt Marécageuse, Bosque pantanoso, "Swamp mixed foret, >30% tree cover, flooded > 9 months" 9, Forest, Gallery forest, Forêt Galerie, Bosque en galería, Riparian forest in valleys or along river edges @@ -87,7 +89,7 @@ In order to properly discern between deforestation and degradation, we require s :header: Deforestation, Degradation "Permanent reduction of forest cover below the forest definition", "A temporary or permanent reduction of forest cover that remains above the forest definition" - "**Conversion of forest** to other land use: agriculture, pasture, mineral exploitation, development, etc.", "Includes areas where timber is exploited or trees are removed, and where forest may be expected to regenerate naturally or with silvicultural methods" + "Conversion of forest to other land use: agriculture, pasture, mineral exploitation, development, etc.", "Includes areas where timber is exploited or trees are removed, and where forest may be expected to regenerate naturally or with silvicultural methods" "Excludes areas of planned deforestation, such as timber extraction, or in areas where the forest is expected to regenerate naturally or with silvicultural methods", "Includes areas where impacts, overexploitation or environmental conditions prohibit regeneration above the forest cover definition" @@ -104,7 +106,7 @@ Deforestation is recognizable in images by a permanent change in forest cover. I Example of degradation """""""""""""""""""""" -Degradation is more difficult to determine because changes are more subtle (sometimes a few trees removed), and tree cover remains above the national definition. It is therefore necessary to look at the whole time series and make sure that the changes are not deforestation. Degradation is also not the same everywhere and will differ by forest type and environmental and human context. +Degradation is more difficult to determine because changes are more subtle (sometimes a few trees removed), and tree cover remains above the national definition. It is therefore necessary to look at the whole time series and make sure that the changes are not deforestation. Degradation is also not the same everywhere and will differ by forest type, as well as environmental and human context. .. thumbnail:: ../_images/workflows/drivers/degradation_example.png :title: Example of degradation @@ -120,7 +122,7 @@ The following date convention was adopted: A product for the year YYYY is considered as of 31 December YYYY. -This convention allows a consistent approach to assessing change products. A change map from **year1** to **year2** will be consistent with both **year1** and **year2** maps. The status of the year takes into account any changes that occurred during the year. +This convention allows a consistent approach to assessing change products. A change map from **Year 1** to **Year 2** will be consistent with both **Year 1** and **Year 2** maps. The status of the year takes into account any changes that occurred during the year. .. _workflows:drivers:drivers: @@ -158,7 +160,7 @@ A total of eight direct drivers were defined by their specific characteristics i * - Industrial forestry - .. thumbnail:: ../_images/workflows/drivers/industrial_forestry.png :group: workflows-drivers - - Large-scale or industrial forestry is recognizable by the presence of logging roads, along which selective logging degradation. These roads may be recent or old, and the canopy can quickly cover them, so all years of imagery acquired over the entire study period are evaluated. + - Large-scale or industrial forestry is recognizable by the presence of logging roads, along with selective logging degradation. These roads may be recent or old, and the canopy can quickly cover them, so all years of imagery acquired over the entire study period are evaluated. * - Artisanal mine - .. thumbnail:: ../_images/workflows/drivers/artisanal_mine.png :group: workflows-drivers @@ -168,7 +170,7 @@ A total of eight direct drivers were defined by their specific characteristics i :group: workflows-drivers - Large-scale mining is characterized by large ponds, open pits and clearings, as well as extensive infrastructure and roads present. -To address the overlapping of drivers in the same location and thus interpret local contexts, our approach identifies archetypes, or common driver combinations which represent realities and processes on the ground. The most common archetype consists of four drivers – artisanal agriculture, artisanal forestry, roads and settlements – which are representative of the agricultural mosaic, or so-called “rural complex”, commonly observed in the region.\ :footcite:`molinario:2020` +To address the overlapping of drivers in the same location and thus interpret local contexts, our approach identifies archetypes, or common driver combinations which represent realities and processes on the ground. The most common archetype consists of four drivers – artisanal agriculture, artisanal forestry, roads and settlements – which are representative of the agricultural mosaic, or so-called “rural complex”, commonly observed in the region (Molinario, 2020). The observed combinations of drivers are grouped into thematic classes or archetypes. @@ -188,15 +190,15 @@ The observed combinations of drivers are grouped into thematic classes or archet Methodology ----------- -The major components of this methodology include the generation of wall-to-wall geospatial data on forest cover types, changes, and discerning areas of deforestation from degradation for the entire central African region. Next, these products are validated via visual interpretation; the presence of various direct drivers are identified to evaluate the direct causes of disturbance, and then interpreted in the context of strategic investments for climate change mitigation and support for national efforts for emission reductions. +The major components of this methodology include the generation of wall-to-wall geospatial data on forest cover types, changes, and discerning areas of deforestation from degradation for the entire Central African region. Next, these products are validated via visual interpretation; the presence of various direct drivers are identified to evaluate the direct causes of disturbance, and then interpreted in the context of strategic investments for climate change mitigation and support for national efforts for emission reductions. The methodology uses FAO’s Open Foris Suite of Tools, including the SEPAL platform, for satellite data analysis, as well as Collect Earth Online (CEO) and Google Earth Engine (GEE). The approach analyses dense satellite time series to generate geospatial data on forest changes, which are then validated and interpreted for direct drivers in five major steps: -#. :ref:`workflows:drivers:mosaic`: processing of optical (Landsat 4/5/7/8) and radar (Sentinel 1/ALOS PALSAR) satellite images to create mosaics for the classification of wall-to-wall maps of vegetation types, recoded to a binary forest mask (following national forest definitions), and forest fragmentation assessment for the baseline year (2015). +#. :ref:`workflows:drivers:mosaic`: Processing of optical (Landsat 4, 5, 7 and 8) and radar (Sentinel-1/ALOS PALSAR) satellite images to create mosaics for the classification of wall-to-wall maps of vegetation types, recoded to a binary forest mask (following national forest definitions), and forest fragmentation assessment for the baseline year (2015). -#. :ref:`workflows:drivers:series`: processing of optical satellite (Landsat 4/5/7/8) time-series data covering 2012–2020 (2012–2015 is the historical time period; monitoring is from 2016–2020), using seasonal models and break-detection algorithms to produce a forest change map for 2015–2020 at the regional scale, identifying areas of both deforestation and degradation. +#. :ref:`workflows:drivers:series`: Processing of optical satellite (Landsat 4, 5, 7 and 8) time series data covering 2012–2020 (2012–2015 is the historical time period; monitoring is from 2016 to 2020), using seasonal models and break-detection algorithms to produce a forest change map for 2015–2020 at the regional scale, identifying areas of both deforestation and degradation. -#. :ref:`workflows:drivers:stratification`: Stratified random sampling is conducted on the change map from Step 2. Systematic validation for all points identified as change, plus a sample of stable points is conducted in Collect Earth Online (CEO), evaluating land cover types, changes and dates of change, as well as the identification of the presence of direct drivers. +#. :ref:`workflows:drivers:stratification`: Stratified random sampling is conducted on the change map from Step 2. Systematic validation for all points identified as change, plus a sample of stable points is conducted in CEO, evaluating land cover types, changes and dates of change, as well as the identification of the presence of direct drivers. #. :ref:`workflows:drivers:quantification`: The quantification of direct drivers by forest types and fragmentation class. @@ -217,9 +219,9 @@ As you can see in this `online animation