Exploring the Potential for Bamboo Agroforestry through a Value Chain Approach in Thailand, using Cloud-based Data Management on Google Earth Engine

doi:10.21203/rs.3.rs-3148270/v1

Download PDF

Article

Exploring the Potential for Bamboo Agroforestry through a Value Chain Approach in Thailand, using Cloud-based Data Management on Google Earth Engine

https://doi.org/10.21203/rs.3.rs-3148270/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The global construction sector consumes almost half of the world's total material production footprint, implying it is the highest single-category material footprint across the global economy. In the search for nature-based alternatives, bamboo grass can become a mainstream building material in the global tropical belt due to significantly shorter rotation times compared to softwood or hardwood species.

To determine bamboo potential within a value-chain work frame, the first step is to evaluate the potential land for cultivation. Technically driven soil selection, with optimal climatic conditions, can generate culms taller than 20 m or, on the contrary, culms of a maximum of 6 to 7 m in height. The correct choice of soil also favors the plantation owner since faster plant development will be observed; the plantation will show a higher number of culms, larger diameters, cheaper production, and shorter times until plant maturity, thus requiring less time to recover the initial investment.

This research presents a remote-sensing-based tool for surveying and exploring bamboo agroforestry potential over the entire national territory of Thailand based on climatic conditions, altitude, topography, existing land cover, and soil characteristics (texture and acidity).

Through the implementation of the research, a total amount of 345,838 km² with viable growth conditions was discovered. 45,968 km² (13.29%) show basic compatibility levels, 242,198 km² (70.03%) show intermediate levels and 57,672 km² (16.68%) present optimal growth conditions. The corresponding regions were located on an interactive geoportal with a 100 m-per-pixel resolution and the ability to benchmark the individual selection criteria.

Earth and environmental sciences/Ecology/Agri ecology

Earth and environmental sciences/Ecology/Tropical ecology

Physical sciences/Materials science/Biomaterials

bamboo agroforestry

remote-sensing

urban management

landscape planning

google earth engine

nature-based materials

1.1 Bamboo and Sustainable Construction

The global construction sector consumes almost half of the world's total material footprint, accounting for the highest material footprint across the global economy (Salonen & Vangsbo 2019). Around 39% of global greenhouse gas (GHG) emissions can be traced to the construction sector, mainly associated with the construction of buildings, but also the production and supply of construction materials (Organization for Economic Cooperation and Development [OECD] 2018). On a finer detail level, the global construction sector uses about 65% of non-metallic minerals, 15% of ferrous metals, and 3% of non-ferrous metals (OECD 2018).

The studies from the Centre for Industrialized Architecture, The Royal Danish Academy, provide comparative quantitative data about the global warming potential (GWP) of standard construction materials, of which ferrous metals values in the range of 1,172 − 28,242 [kg CO₂ eq / m3], concrete and clays are lower with 123–898 [kg CO₂ eq / m3]. In contrast, biomass-based materials have a negative effect ranging from − 1,063 to -128 [kg CO₂ eq / m3], which denotes their capability for carbon sequestration and storage (Kauschen et al. 2020).

As a biogenic alternative, bamboo could become the predominant building material to mitigate GHG. Especially its shorter rotation times compared to softwood or hardwood species (van der Lugt & Vogtländer 2015). One hectare (ha) of bamboo can produce almost five times as much biomass (up to 19,800 kg/ha) and simultaneously bind almost three times as much carbon (927,630 kg/ha) as a traditional softwood plantation (Dieren 2020).

It is also noteworthy that in most areas of the developing world, new urban expansion takes place over the most fertile lands, thereby creating competition between land use for construction and agriculture (International Resource Panel [IRP], 2017). Under the basic premise of regarding urban construction as a potential carbon storage infrastructure at a global scale, the detailed analysis of GHG emissions sequestration capacity of bamboo-based products becomes of critical importance. This is particularly relevant along tropical/subtropical regions, where bamboo naturally grows, and higher building material demand for construction in urban environments will be the greatest over the next 30 years (Churkina et al. 2020). In such contexts, bamboo farming can contribute to reducing deforestation and generate a reliable income source for small-scale landowners in rural communities (Churkina et al. 2020).

Previous research on industrialized bamboo carbon storage seems to coincide around an average figure of 693 kg of carbon stored per 1,000 kg of the finished product (higher than the 480 kg calculated for wooden products), however, the range goes from a minimum of 674 kg to a maximum of 707 kg (Verspeek & van der Burgh 2017; van der Lugt & Vogtländer 2015) depending on multiple factors, such as the specific bamboo species, the manufacturing process, and the end-product characteristics These calculations account for biogenic CO₂ (above and underground), land-use change, processing losses, and market growth (Vogtländer et al. 2014). One of the available commercial bamboo product life cycle analysis (LCA) based on EN-15804 standard (CEN/TC 350, 2013) for Environmental Product Declarations (EPD), considered within the system boundary: (1) production stage, (2) construction stage, (3) Use stage, (4) end-of-life stage, rendered the following results in Table 1. This analysis is based on the bamboo harvested and processed in southern China and commercialized in the European market. (Verspeek & van der Burgh 2017)

Table 1

LCA results for the environmental categories on various bamboo products (per kilogram of finished product).
End-use	Product density (kg/m³)	Bamboo content (%)	Global warming potential (kg CO₂ eq.)	Ozone layer depletion (kg CFK-11 eq.)	Total energy consumed (MJ)	Water consumed(m³)	Waste (non-hazardous) (kg)
Indoor paneling and furniture	700	97	0.286	1.476E-08	0.407	0.835	113,039
Indoor semi-structural	1,080	95.5	0.261	6.36E-09	1.607	0.854	113,036
Outdoor semi-structural	1,200	93	0.208	6.75E-09	0.990	0.790	110,924
Average	993	95.2	0.252	9.27E-09	1.001	0.826	112,333

1.2 Bamboo Agroforestry

Considering the LCA, the phase of building life cycle products starts with the production stage, where the raw material supply is the initial substage (CEN/TC 350, 2013). The bamboo agroforestry planning is the first stepping stone toward the GHG mitigation capacity of bamboo-based products. Moreover, bamboo agroforestry contributes to ecosystem health improvement and sustainable economic growth.

From an ecological perspective, bamboo is a grass species with a fast-growing speed and has a vast connected stems root system underground (Shen et al. 2019). These characteristics result in a high annual yield, leading to corresponding high carbon sequestration. Especially the bamboo agroforestry systems, combined with other compatible plant species. Up to 58 species of shrubs, 35 species of birds, and 28 species of broadleaf trees have been recorded to be compatible with healthy bamboo growth and accounted for significantly less forest damage under extreme weather conditions (Zheng & Hong 1998). This combination would be a fast-acting, highly effective pathway for climate change alleviation and loss of biodiversity (van der Lugt & Vogtländer 2015; Yiping & Henley 2010).

From a socioeconomic standpoint, the rotation cycle of bamboo is around one year, after a 4–5 years initial development period. This could generate a reliable and steady income, compared to the longer cycles of wood production (van der Lugt & Vogtländer 2015). Annual harvesting allows for bamboo incorporation within the existing economic crops of tropical and subtropical regions, as well as for more biodiversity and agricultural diversity, thus decreasing the risk of price fluctuation. Furthermore, the implementation of bamboo as a building material would provide the local labor market with more job opportunities at several points, i.e., pre-processing, treatment, manufacture, and installation.

1.3 The Value Chain Approach

To determine the complexity of potential improvement for bamboo as a building material which is relevant to lessen GHG emissions and land use dysfunctionalities with their corresponding deforestation impact, local economy, and biodiversity loss, the value chain approach (VCA) could be used to gain a better understanding of its material sourcing towards circularity and identify ways to improve its overall efficiency and effectiveness,

In 2021, the United Nations Environment Program (UNEP) published the first report on the VCA, partially based on UNEP’s “Eco-Innovation Approach” and previous work from the “Life Cycle Initiative,” underlines the importance of a systemic vision regarding the connection between natural resources use, its environmental impact, and the socioeconomic reality of existing production and consumption practices (United Nations Environment Programme [UNEP] 2021).

Departing from the UN’s definition, a value chain (VC) can be understood as the group of activities that provide or receive value from designing, making, distributing, retailing, and consuming a product. As illustrated in Fig. 1, this includes the extraction and provision of raw materials, and the product´s processing after its useful service life. The VCA targets to engage all actors along the life cycle, to identify the drivers and barriers that shape its operations, and to reveal the most promising solutions for identified problem hotspots (Lopez y Royo et al. 2021).

This approach relies on data analysis and discussion organized in three steps: (1) understand the VC and identify hotspots; (2) consolidate existing action and identify opportunities to address the located hotspots; and (3) define a common agenda and prioritize through a participatory process. This research project corresponds to sub-step 1.b: Map data on natural resource use and impacts to the stages of the VC. A closer look into the analysis of the construction VC concludes that the largest natural resource consumption and environmental consequences occur during the material manufacturing stage. The most significant participants in the VC are governments, international organizations, financial institutions, and big market companies, which primarily operate during the financing and planning stages. The activities along the remainder of the VC are mostly shaped by the decisions made in these phases (Lopez y Royo et al. 2021).

Considering that traditional construction material manufacture uses a combination of nature-sourced inputs, the predominating dependence on virgin concrete and metals within the construction VC produces negative environmental impacts over all the following aspects: (1) deforestation and land-use changes; (2) biodiversity loss; (3) water scarcity and pollution; (4) soil pollution and run-off; (5) air pollution; (6) GHG emissions (Environmental Coalition on Standards [ECOS] 2020). A publicly endorsed bamboo agroforestry strategy aimed at the manufacture of modular building components could directly cut down the negative impacts of aspects 1, 2, and 5, while indirectly contributing to aspects 3 and 4.

1.4 Potential vs. Existing Bamboo Area

Geographic information systems (GIS), remote sensing (RS), and spatial data processing from satellite sources constitute a valuable tool for evaluating the contribution of different geo-climatic features in the identification of agricultural stands, and the potential determination of viable production lands for new forestry resources. (Li et al. 2016; Qi et al. 2018; Zhao et al. 2018).

Current research about bamboo agroforestry areas can be classified into two main types, (1) to investigate the existing condition (surveying), (2) to find new potential farming sites according to multiple growth conditions mapped in geospatial data. Existing bamboo areas are often analyzed using RS imagery from satellites in combination with aerial photography taken from traditional aircrafts or drones, which have been developing since the late 1990s. These techniques can assess the global bamboo coverage and even distinguish the phenology of bamboo with the improvement of the technology over time (Dida et al., 2021). A relatively recent breakthrough of this remote imagery interpretation is enhanced by the advance of machine learning (ML) classification algorithms over the last ten years, which significantly increased the accuracy rate as shown in spatiotemporal dynamics of bamboo forests analysis from China (Qi et al. 2018; You et al. 2020).

This research, however, is focused on the latter type and constitutes an attempt to develop a methodological protocol to find suitable areas for future bamboo cultivation. The creation of such baseline mapping can serve as the departure point for new land-use policy creation in the pursuit of at least three clear outcomes: (1) transition from a resource extractive tendency toward prospective land regeneration, (2) establish bamboo-based small-scale industries along the rural landscape, (3) create a carbon trading mechanism supported by the carbon sequestration and storage capacity of the fast-growing bamboo biomass (Devi & Singh 2021).

Based mostly on GIS, a similar technique was applied as early as 2010 for the potential bamboo research in Mozambique, which uses topography, physiography, climate, and infrastructure layers to create output maps such as the optimal location for bamboo-based charcoal plant set-up, or suitable areas for bamboo nurseries establishment (International Union for Conservation of Nature [IUCN] 2010).

That process worked on numerous geospatial datasets collected from public access or state authorities to evaluate several earth surface conditions leading to the evidence-based prediction of feasibility for new potential bamboo agroforestry sites. However, the processing capacity limitations of personal computers and small-scale clusters pose a barrier to multi-phase RS imagery analysis, which is better handled by cloud-based distributed computing architecture such as the Google Earth Engine (GEE) platform (Qi et al. 2018). This analysis strategy has been successfully implemented in nationwide bamboo mapping research along the east African region, specifically in Ethiopia, Kenya, and Uganda (Zhao et al. 2018).

As the scale of the present research covers the entirety of Thailand´s national territory, the analyzing process was conducted through GEE, given its ability to manage big data volumes much more efficiently than a desktop application as well as its potential for global scalability in upcoming research iterations. Lastly, the use of GEE enables the results of this research to be explored through an online access platform with the final map for simple interaction with the output data.

2.1 Scale: From Thailand to Global

Thailand is located in Southeast Asia, at the center of the Indochinese Peninsula, with a total area of 513,120 km² (Fig. 2). Most of the country's climate is classified as Köppen's tropical savanna, with some regions having a tropical monsoon or tropical rainforest climate. (Beck et al. 2018) The annual average temperature is 27.1°C with an average total annual rainfall of 1,587.7 mm (1981–2010) (Thai Meteorological Department, Ministry of Digital Economy and Society [TMD] 2021). Bamboo is one of Thailand's most socio-economically important species, generally found in natural forests and developed for cultivation (Ramyarangsi 1985).

Historically, bamboo was harvested as food, household handicrafts, source materials for construction (which remained in the vernacular architecture), especially in the northern provinces (Correia et al. 2013). The research selected Thailand as the testbed to trial the experiment due to its potential for bamboo production, the country´s scale, and its cultural compatibility aspect. Further research could be scaled up or down, mosaicking the source data as a detailed area on the municipal level, or globally analyzed on a planetary scale. The proposed methodology can be implemented to visualize entire material flows and identify development opportunities along the VCA on any preferable scale.

2.2 From Bamboo to Other Crops

The research analyzed the potential for bamboo growth and harvesting based on critical natural factors influencing agriculture – precipitation, soil, and topography based on the specific bamboo species requirement, in this case, Dendrocalamus Asper due to the high yielding rate and its endemic nature over the study area. Each parameter defines the possible area layer for individual growth aspects and is then aggregated with the others to define the areas containing valid data for all criteria. This methodology can be replicated to another bamboo genus, depending on the geographical condition, or expanded further to other economically valuable crops such as cereals, oil-yielding crops, or lumber trees by defining different parameter ranges (filters) on the geospatial data, as shown in Fig. 3.

To explore the initial steps in new material VC and identify their potential in extensive heterogeneous landscapes, however, it further requires specific criteria for different vegetation types.

2.3 Datasets

Data Source Description. The geospatial data for this research was gathered from two primary sources: (1) geo-referenced raster images generated from the public data catalog for digital elevation model (DEM) and bioclimatic map in the global scale dataset, combined with (2) local vectorized data provided by the Land Development Department (LDD) and Royal Forest Department (RFD) under the Government of Thailand, who generated it using a combination of remote sensing and physical surveyed on-site validation.

The soil properties data contained the description in Thai soil categorization with 62 types which were aggregated with the texture and acidity (pH level) data according to LDD soil literature supplements (Udomsri & Hoontrakul 1997). Some areas are missing LLD soil properties data, such as gullied lands or areas with a slope over 35% (19.29°), as the area was difficult to manage or maintain for agriculture, which aligns with this research's objective. Land use and land cover (LULC) with level-2 classification are divided into 32 types. To achieve a country-wide dataset, 76 tiles with every region in Thailand were combined using QGIS 3.18 and its native vector merging algorithm. The same method was used on the existing oil palm dataset to aggregate five region files. Lastly, the forest cover map contains 17 types, including existing bamboo forests. The polygon shapefiles were organized in QGIS and exported as shapefiles to upload into the GEE system. The vector files were rasterized and exported as raster assets for web-based optimization with a resolution of 100 m per pixel. The entire list of input data included in the workflow is presented in Table 2.

Table 2

Input data for the interpretation of potential bamboo agroforestry land.
Input data	Name of dataset	Year	Data type	Resolution (m)	Provider
DEM	NASA SRTM Digital Elevation 30m	2000	Raster	30	NASA / USGS / JPL-Caltech
Bioclimatic map	WorldClim BIO Variables V1	1991	Raster	927.67	University of California, Berkeley
Soil properties map	Soil Series	2019	Vector	50	LDD
LULC	Landuse Thailand – Level 2	2018	Vector	12.5	LDD
Forest map	Thailand forest type database	2018	Vector	25	RFD
Oil palm map	Zoning by Agri-Map, Zoning Palm	2019	Vector	12.5	LDD

Driving Parameters Selection. Agricultural production is affected by various social and ecological factors, which are often challenging to integrate into the analytical framework. This research implements the capacity of online geospatial data analysis processed on a cloud-based system that could potentially manage petabytes of information to aggregate and weight geo-climatic conditions that generate a comparable score and identify the potential hotspots for bamboo farming. The crucial geo-climatic factors for agriculture include (1) terrain slope, (2) terrain elevation, (3) soil acidity, (4) soil texture, (5) average annual rainfall, (6) temperature. However, the research team decided not to include this variable due to Thailand´s geographic location, close to the Equator, with a stable average temperature (22–28°C) all year round. For the existing land cover factor, LULC and forest locations were used to consider the national economy's land resources and land use types aside from bio-physical systems. The land distribution information provides the specific scope that could be exploited for the potential bamboo agroforestry area. Interested stakeholders can use this information in decision-making to eliminate incompatible land uses, such as high-density urban clusters, humid evergreen forest, or sensitive wetlands. Lastly, the research incorporated the oil palm cover, resulted as one of the fastest-growing agricultural land uses. Its expansion significantly impacts the environment in a negative way; deforestation lessens the habitat area of endangered species and degrades land quality (Meijaard et al. 2020). The oil palm parameter underlines the potential benefits of significantly substituting mixed bamboo agroforestry to achieve higher ecosystem biodiversity, better soil quality, more resilient agricultural land, and a more sustainable cultivated development (Yiping et al. 2010).

Acceptable Thresholds. The data input was initially classified into viable and non-viable values (binary). To define the viable values, the developed workflow focuses on framing the ideal growth conditions based on the agricultural requirements of various high-yield bamboo species (Phyllostachys edulis, Dendrocalamus Asper, Guadua Angustifolia), thus filtering the suitable land surface for agroforestry according to:

a. Terrain slope – several studies show a wide-range optimal slope for bamboo; however, the highest productivity is found around the 25-degree slope (Cheng et al. 2015). Another indicator is the Thailand forest type database, which shows a similar result with bamboo forests situated on a range from flat land to hillsides with over 35-degree slopes. However, areas with more than 20% slope have a high erosion risk and difficult accessibility for farming (Jarasiunas 2016), reason for which this becomes the maximum admissible slope value for this research.

b. Terrain elevation – the studies on bamboo productivity based on elevation show bamboo could be found at altitudes as high as 7,000 m in rare cases (Shi et al. 2020). However, the elevation with the highest density of existing forests falls between 0–2,400 m above sea level (Schröder 2021). Over this altitude, lower air temperatures start to play a limiting factor.

c. Soil acidity – the suitable pH condition falls in slightly acidic soil to neutral, the range from several pieces of research are shifting within pH values of 4.5-7 (Fu 2001; Da Silva Sobrinho Junior et al. 2009).

d. Soil texture – bamboo can grow in different soil classifications. However, the smaller-size particles of soil, such as heavy clays, are less suitable for bamboo cultivation due to the high density that restricts the rhizome and root growth. While the too-light texture, such as sand, has an inadequate water-holding capacity (Kleinhenz & Midmore 2001). The acceptable threshold in this research is limited to combining loamy soils with other textures, as defined in the USDA Textural Soil Classification: sandy loam, loam, clay loam, and silty loam (United States Department of Agriculture [USDA] 1987).

e. Average annual rainfall –in Thailand, the minimum water requirement for bamboo growth was reported to be 1,000 mm/year without the maximum indicated, yet several studies demonstrated the yearly maximum admissible rainfall within 4,000–5,000 mm, which has been defined as the threshold limit of average annual precipitation for this research (Kleinhenz & Midmore 2001).

f. LULC – LDD classifies the LULC level-2 data into five main categories: (1) urban and built-up land, (2) agricultural land, (3) forest land, (4) waterbody, (5) miscellaneous land, with 32 sub-types. The data were analyzed according to the possibility of developing the area for future bamboo cultivation land. The acceptable threshold is shown in detail in Table 4.

g. Forest cover – the entirety of existing forest areas, wetlands, and natural land cover is considered as non-viable values to preserve fragile ecosystems and their biodiversity and mitigate climate change.

h. Oil palm – as previously mentioned regarding the climate impact of oil palm production, this intensive agricultural area represents an additional viable layer to promote as the prioritized land cover shifts towards bamboo production after its productive lifespan is over (circa 20 years).

2.4 Calculations

Scoring System. Multi-variable analysis plays a crucial role in land development policy. After gathering geospatial data related to bamboo farming factors and literature reviews, the previously binary filter was created; furthermore, this study proposes a scoring system to benchmark the potential for bamboo cultivation development, corresponding to the VCA. The system classifies numeric or non-numeric data in a range from 0–3 (normalization) with the criteria as shown in Table 3. However, due to limitations on the available data inputs, the research could only apply this scope of the method to ecological factor parameters (a-e). The scoring approach offers an intelligible area comparison for complex data; therefore, decision-makers, investors, or general audiences could effortlessly interpret the analyzed result and use this information through corresponding actions. The score value was added to the geospatial data by filtering and overwriting selected pixels' data properties.

Table 3

Scoring system on ecological factor parameters.
Layer name					Reference
	0	1	2	3
Terrain slope (degree)	> 35	0–5, 25–35	5–15, 20–25	15–20	Cheng et al. (2015), Jarasiunas (2016)
Terrain elevation (m)	> 2,400	1,800-2,400	1,200-1,800	0–1,200	Shi et al. (2020), Schröder (2021)
Soil acidity (pH)	0-4.5, >7	4.5-5, 7-7.25	5-5.25, 6.5-7	5.5–6.5	Fu (2001), Da et al. (2009), Government of Western Australia (2014)
Soil texture	sand, clay, silt	silty clay loam, sandy clay loam	-	sandy loam, loam, clay loam, silty loam	Kleinhenz & Midmore (2001), Nath et al. (2015)
Average annual rainfall (mm)	0-1000, > 5000	1000–1500, 4000–5000	1500–2000, 2500–4000	2000–2500	Kleinhenz & Midmore (2001), Schröder (2021)

This research did not create a specific weighting coefficient for each factor to generate the formula for the final data score due to the lack of consensus on the matter from multiple literature sources; for example, two pieces of research indicate the contradiction in the slope parameter importance. One suggests slope as the most crucial variable for enhancing bamboo productivity (Cheng et al. 2015), while another clearly states slope aspect had no significant effect (Chen et al. 2017).

Masking. The masking step was included to differentiate viable values (1–3) from non-viable ones (0), as the output from the scoring system step still contains scores with unsuitable areas. The masking process generates an output image that retains the metadata and footprint of the input image. Pixels will be excluded where the mask changes from zero to another value. The filter input variable is a mask image containing only the viable value area pixels. It is noteworthy to mention that the input for this step must be raster and not vector data. In vector data such as the soil texture properties, the workflow requires an additional rasterization process, which was achieved through GEE JavaScript API, as shown in Fig. 4.

Map Superposition. Overlaying is the last step to generate the final map to identify the potential bamboo agriculture hotspots through the geographical intersection of previously mentioned databases. The overlaying map composites all the images containing the same data attribute (score) to a single image representation, using the mean value as the data reducer for each pixel. After generating this mean map, it must be masked to exclude the pixels with a non-viable value. The map superposition shows area data as every pixel shown on this map can grow bamboo according to all geo-climatic parameters and contains a comparable score system (Fig. 5).

Qualitative Layers. Non-numerical layers (strings) that are by definition not compatible with the scoring system will be used as binary filters (true or false). In this case, this method applies to social structure factor data: (1) LULC; (2) forest map; (3) oil palm map. These layers are used in two ways, excluding the non-viable area or additionally promoting specific areas, as shown in Table 4. The meaning of qualitative layers differs from the scoring layers; the data would not be included in the score calculation but instead, operate as extra information layers for decision-making.

Table 4

Binary filter system on LULC parameters.
Layer name
	false (non-viable)	true (viable)
LULC	A8 Aquatic plant A9 Aquacultural land F1 Evergreen Forest F2 Deciduous Forest F3 Mangrove Forest F4 Swamp Forest F5 Forest plantation F6 Agroforestry F7 Beach Forest U1 Urban and Commercial area U2 Residential area U3Governmental and Institutional land U4 Transportation Communication and Utilities U5 Industrial land U6 Other Built-up areas U7 Golf Course W1 Natural water body W2 Artificial water body M3 Mine and Pit M4 Other Miscellaneous lands M5 Salt flat M6 Beach	A0 Integrated farm A1 Paddy field (Rice field) A2 Field crop A3 Perennial A4 Orchard A5 Horticulture A6 Swidden cultivation/ Shifting cultivation A7 Pasture and Farmhouse M1 Rangeland M2 Mash and Swamp
Forest map	All Area	-
Oil palm map	-	All Area

In further research, the qualitative layers could be included in the scoring system to enrich the benchmarking output with adequate literature support; for instance, the LULC or forest class could be classified according to local land development guidelines.

2.5 Validation

This research compares predicted data to the existing current RS imagery from official authorities to validate the prediction outcome. The most up-to-date information on the current bamboo area in Thailand was found in LULC level-3 data conducted between 2018–2021 at a scale of 1:25,000 provided by LLD from the LANDSAT 8 OLI satellite. According to the data description, the classification accuracy ranges between 85 and 98%, depending on the city. The datasets are separated into 232 subclasses, with bamboo falling under subclasses A315 (bamboo) and M103 (giant thorny bamboo).

The employed procedure filters the existing bamboo area (LULC l-3 vector data) using the predicted potential bamboo agroforestry layer (the research result as raster data, vectorized at a scale of 1km/pixel) as a clipping mask. The area of the filtered land surface is calculated with the “.geometry().area()” command and printed out with “ee.Number().divide(1e6)” to convert the result into km² using the GEE interface. Finally, the result of this operation is divided by the total existing bamboo area from LULC l-3, to determine the percentage of the masked area that corresponds to the actual data (i.e., where the prediction matches RS observations).

Due to availability limitations, several cities lacked data, particularly in the country’s central region. The complete sets of data are available only for the northeast and south regions. As a result, the described validation method could only be performed on these two specific regions, which contain 34 from 77 different cities and cover 242,702 km² (47% of the nation's surface).

3.1 Final Overlay Map

Individual Layers.

Terrain Slope. The slope is one of the factors determining the location of crops, as a higher sloped area requires significantly more resources for human access or mechanical harvesting systems. Regarding natural factors, the slope could positively affect the quality of daylight, nutrients, and water access for the plants. Therefore, the slope is vital in identifying and evaluating suitable areas for bamboo nurseries and plantations. Figure 6 shows that 98% of the land in the study area is capable of bamboo farming in terms of the slope parameter. However, more than two-thirds of the suitable land fall into the score one range (1 point out of 3), as shown in Table 5.

Table 5

Terrain slope result.
Data input	Score	Criteria (degree)	Output area (km²)
DEM Band: slope	0	> 35	8,812
	1	0–5, 25–35	349,689
	2	5–15, 20–25	118,641
	3	15–20	35,978

Terrain Elevation. Conventionally, altitude would impact environmental factors such as temperature and light, thus affecting plant growth. Most bamboo species are native to warm and moist climates and have a growth boundary line on the high elevation level of over 2,400 m (Shen et al. 2019). On average, the land in Thailand is at a comparatively low elevation, about 286.44 meters above sea level (Farr et al. 2007). As a result, the analysis of the elevation factor shows a score of three (3 points) on 506,424 km², which is distinctly suitable for bamboo cultivation, as shown in Table 6 and Fig. 7.

Table 6

Terrain elevation result.
Data input	Score	Criteria (m)	Output area (km²)
DEM Band: elevation	0	> 2400	5
	1	1800–2400	130
	2	1200–1800	7,274
	3	0-1200	506,424

Soil Acidity Soil pH measures the acidity or alkalinity state found in the ground, directly influencing plant growth through factors such as soil bacteria, nutrients, and structure (Kleinhenz & Midmore 2001). Bamboo prefers slightly acidic to neutral soil, where the range could be varied from 4.5 to 7, but several researchers and guidelines specifically recommend a sharper range from 5.5 to 6.5 (IUCN 2010). Figure 8 illustrates a wide range of soil pH in the study area. More than 66,280 km² of areas are unsuitable for bamboo, as shown in Table 7. Most of the land fits the viable criteria, and the highest score is found mostly in the northern regions of Thailand.

Table 7

Soil pH result.
Data input	Score	Criteria (m)	Output area (km²)
Soil properties map property: ‘pH’	0	0-4.5, >7	66,280
	1	4.5-5, 6.75-7	185,562
	2	5-5.25, 6.5–6.75	61,114
	3	5.5–6.5	200,877

Soil Texture. Soil texture influences aeration, water movement, nutrient storage, and erodibility. The high-density soil restricts the bamboo rhizome and root growth, while the sandy texture has low water capacity, so the analysis focuses on finding the mixed soil texture as loam (Nath et al. 2015). The soil texture parameter excludes the vast amount of land in the central part of the country (Fig. 9), where soil types mainly consist of heavy clay texture shown in Table 8.

Table 8

Soil texture result.
Data input	Score	Criteria	Output area (km²)
Soil properties map Column: ‘Texture’	0	sand, clay, silt	134,563
	1	silty clay loam, sandy clay loam	82,574
	3	sandy loam, loam, clay loam, silty loam	296,696

Average Annual Rainfall. Water availability is a vital parameter to any agricultural production, severely impacting farm yield. The preferred average annual rainfall range shifts slightly on each species. Still, the minimum tolerance of around 1,000 mm/year is the bottom threshold (IUCN 2010), which complies with the local bamboo cultivation guideline from Thailand´s Department of National Parks, Wildlife and Plant Conservation. Most of the study areas are suitable for bamboo farming, as shown in Table 9. Additionally, the southern regions contain the highest score for the average yearly rainfall factor (Fig. 10).

Table 9

Annual rainfall results.
Data input	Score	Criteria (mm)	Output area (km²)
Bioclimatic map Band: ‘bio12’	0	0-1000, > 5000	7,753
	1	1,000–1,500, 4,000–5,000	334,274
	2	1,500-2,000, 2,500-4,000	116,217
	3	2,000–2,500	55,589

LULC. The LULC shows the physical distribution of how people utilize the earth’s surface for socio-economic activities, including nature-based and artificial features. The data plays a primary role in land planning and management. A VCA to articulate bamboo cultivation, development, and processing would heavily depend on these conditions. This research uses this parameter to filter areas that could be turned into bamboo farming without causing drastic changes in the natural landscape or the urban fabric by focusing on sites currently occupied with agricultural fields or empty land, as shown in Table 10. The resulting illustration (Fig. 11) exhibits the masking layer binary filter to avoid unsuitable areas after going through the scoring system.

Table 10

LULC result.
Data input	Binary filter	Criteria	Output area (km²)
LULC Column: LUL2_CODE	true	A0, A1, A2, A3, A4, A5, A6, A7, M1, M2	234,485
LULC Column: LUL2_CODE	false	A8, A9, F1, F2, F3, F4, F5, F6, F7, U1, U2, U3, U4, U5, U6, U7, W1, W2, M3, M4, M5, M6	282,992

Forest. Forest cover is critical to decelerate climate change, capturing and sinking CO₂ for medium to long periods. They are also vital for healthy ecosystems and biodiversity conservation. The woodland offers a habitat for living creatures, prevents soil erosion, and preserves biological cycles, especially in tropical regions with high species richness (Churkina et al. 2020). Therefore, existing forest surfaces will be considered unsuitable values, as shown in Fig. 12, based on conditions according to Table 11.

Table 11

Forest surface result.
Data input	Binary filter	Criteria	Output area (km²)
Forest Column: ftype_eng	true	Non-forest area	353,665
Forest Column: ftype_eng	false	Forest area	163,981

Oil Palm. This final masking layer (Fig. 13) shows a map of Thailand's existing oil palm plantations. Several authors coincide on the high-environmental impact of oil palm and its currently uncontrolled expansive trend all over Southeast Asia and Central America (Descals et al. 2021). Promoting bamboo agroforestry, which is widely considered more beneficial for ecosystem biodiversity species. Considering this, all the area currently covered by oil palm (Table 12) is considered to be viable land for crop substitution. It is added to the LULC layer after undergoing the scoring system.

Table 12

Oil palm result.
Data input	Binary filter	Criteria	Output area (km²)
Oil palm	true	Oil palm	7,413

Scoring System. After mosaicking, compositing, and masking the individual layers, the final result combines the scoring system with the qualitative layers, as shown in Fig. 14a. First, the development of the mean-score map illustrates the area containing viable scores 1–3 in all geo-climatic parameters; the higher the mean score indicated, the higher potential for bamboo cultivation. If a specific pixel includes a value of 0 in any single parameter, it is automatically excluded from the final map through the masking process. The advantage of this scoring technique is that each pixel retains the data from its individual layers, which can be retrieved upon the interactive user interface (UI) and generate a basic scoring plot for better understanding the complex condition of that specific focus area. (Fig. 14b.)

Second, the qualitative layers were respectively subtracted and added over the mean-score map as a binary filter to enrich the aspect of social structure parameters. The LULC and forest layers are used as an “unsuitable land-use filter” (false) to demote the masked area from the final map. The oil palm filter serves as an additional layer to promote substituting environmentally damaging activities with new bamboo agroforestry production. The two systems are displayed with different plots to clarify the interpreted data's different nature (quantitative/qualitative).

3.2 Data Validation

After obtaining the final overlaid map as the predicted potential bamboo area for agroforestry, the layer was flattened and vectorized at a resolution of 1km/pixel, which correlated with the largest scale of the input dataset (existing bamboo reference layer) to be masked during the comparison and validation process. The calculated area result is shown in Table 13.

Table 13

Validation area result
Data input	Southern region (km²)	Northeast region (km²)	Total bamboo area (km²)
Existing bamboo area	1.0478	35.3479	36.3957
Existing bamboo area within the mask	0.7993	33.3056	34.1666

The existing bamboo area filtered within the predicted potential bamboo area mask is 34.167 km² or 93.88% of the total existing bamboo area. This coincidence percentage shows the validity of this prediction method, especially when considering the accuracy of the original LULC dataset at 85–95%.

3.3 Online App

GEE provides the possibility to create an online interactive app that can be accessed by a user interface (UI). The interface helps with data exchange between the final geo-database and the users’ areas of interest. It includes simple input widgets like view-control buttons and geocoding textboxes and more complex commands like chart generation from the given coordinates or recalling the data within a custom polygon for further calculation.

Layer Control and Visibility. The layer manager is situated in the upper right corner of the application (Fig. 15). The widget allows the user to adjust the display of all the geo-climatic, and land cover layers available in the app. Users can move the sliders to adjust layers’ opacity or toggle the visibility for individual layers, thus creating interactive visual configurations to display parameters according to the field of interest.

The Inspector Tools. An inspector widget is a benchmarking tool for comparing data pixel by pixel (100 m resolution). The panel on the bottom left contains the basic instructions, selected latitude-longitude coordinates, and chart plots. The charts present the individual geo-climatic parameters score and an overall mean score for the clicked location. The function is set to regenerate on every user’s click. According to the value chain approach, the inspector tool is the key to identifying the drivers or barriers and revealing the possible new bamboo hotspots.

4.1 Availability of High-resolution Satellite Imagery

Two key factors influence the technical requirements of the input databases: (1) image resolution (2) age of the collected data. The input on this research is based on a mix of remote sensing data and locally sourced/validated databases. The former involves two-dimensional processing and geostatistical methods, including reclassification or thresholding. Before aggregating the data, the scale of analysis must be identified as low-resolution imagery would lead to misinterpretation in the fine-scale analysis; at the same time, an unnecessarily high resolution would significantly decelerate data processing speed, particularly in this case where interactive data requests on every user interaction (front-end) would imply a heavy load on the cloud-based service. For example, a 1 km scale would suit the global or regional analysis, however, when it comes to a local or peri-urban frame, one pixel represents over 1 km2 area which is too broad. In contrast, less than 30 meters scale is weighty to process for the regional scale, especially for vector data. Eventually, the analysis script must contain the data reducer or maximum pixel limitation to prevent a constant error in the UI.

Secondly, the year of data collected affects the accuracy as the earth’s surface constantly changes. In some cases, older bioclimatic maps would be acceptable, as the data requires 14-year averages to identify consistent climate patterns. Nevertheless, the parameters vary faster based on anthropological activities and global warming, as seen in agricultural lands or forest areas shifting annually. When these layers overlay others from different years, the result might slightly differ from reality, therefore, a similar year (within a 5-year range) of origin for the collected data is highly recommended.

Lastly, in the data access aspect, GEE provides a wide range of public datasets, such as Sentinel-2 or Next Generation Weather Radar (NEXRAD), that can be seamlessly accessed and integrated into the applications. However, more detailed datasets sometimes need to be sourced through the local authorities. In this case, the local databases come directly from an offline repository, which requires GIS pre-processing for data preparation or conversion before its upload to the online platform.

4.2 Different Species with Different Thresholds

The geospatial analysis through coding and the use of statistical algorithms opens the possibility of further adjusting the tool based on specific species selection. The parameters can be quickly adjusted on the coding script by editing the threshold range value and re-compiling. As mentioned, this research determined its focus on the native bamboo species previously existing in the study area. In other cases, these criteria could be based on indigeneity, or physical-mechanical characteristics, for instance, “morphology, density and dimensions of bamboo fibers” (Rusch at el. 2019). The subfamily of bamboo is divided into three tribes: Arundinarieae (temperate woody), Bambuseae (paleotropical and neotropical woody), and Olyreae (the herbaceous) (Kelchner 2013). However, the latter subfamily lacks a woody stem. Table 14 shows possible parameter adaptations of the selected bamboo species commonly used for building and construction.

Table 14. Scoring system on ecological factor parameters.

4.3 Uses for Policymaking

As mentioned before, agricultural production is affected by various social and ecological factors, which are often challenging to integrate into an analytical framework.

This research adopts the capacity of geospatial data analysis on a cloud-based system that could manage petabytes of information to consolidate geo-climatic, socio-economic, and georeferenced conditions to generate a comparable score that identifies the potential hotspots for bamboo farming. By doing this, the corresponding institutions offer a technically validated platform for political decision-making on the matter.

The project´s most relevant findings are summarized in Table 15 as follows:

Table 15

Final land availability results.
Data input	Score group	Mean Score range	Output area (km²)
Integrated conditions for Bamboo agroforestry development	0	Unviable	167,995
	1	1.0–1.5	45,968
	2	1.51–2.5	242,198
	3	2.51–3.0	57,672
Total viable land	N-A	1–3	345,838
Total country surface	N-A	N-A	513,833

It appears to be clear that most of the viable land tends to display mean score values close to the score 2.0 (242,198), with an even geographical distribution across the national territory. The second largest group corresponds to land with mean scores close to 3.0 (57,672), which tend to be found in the country’s southern regions. Finally, the least favorable, yet still viable, lands are found in the northern regions, and display mean score values close to 1.0 (45,968 km2).

Sustainable Land Development.

The benefits of sustainable bamboo agroforestry have been argued by several authors over the last two decades (Verkerk et al. 2021; Dida, J. J. V. et al. 2018; Xinzhang et al. 2011). Furthermore, the potential for bamboo products articulation into the construction industry has a proven high potential for atmospheric carbon sequestration and overall climate change mitigation (Churkina et al. 2020; Shen et al. 2019; van der Lugt & Vogtländer 2015).

This research project and its associated outputs propose a shift from a traditional approach of existing resource discovery (mapping) towards a more proactive technical feasibility analysis. This way, the focus is transferred from altering existing ecosystems to retrofitting already modified landscapes and even recovering damaged lands from intensive human exploitation.

This conceptual shift constitutes the first necessary stage in the process of ideating, planning, designing, and implementing a new nature-based value chain for the construction and associated industries with a clear aim for environmental sustainability and economic circularity.

As a final map, the output database clearly shows the regions most likely to host highly productive bamboo agroforests, which could potentially amount to a gross volume of 6.92E + 11 kg (raw), 228,253,080 m³ finished product, and 4.79E + 11 kg of captured CO2eq (Song et al. 2011).

Ethics approval. “Not applicable”

Competing interests. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Authors' contributions. All authors contributed substantially to the study’s conception and design. Material preparation, data collection, and analysis were performed by [Wasita Amatyakul] and [Sebastian Orozco]. The first draft of the manuscript was written by [Wasita Amatyakul] and [Sebastian Orozco] on an approximately 6:4 proportion ratio. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Funding. The participation of author [Sebastian Orozco] in this project was supported by the German Academic Exchange Service (DAAD); and the Costa Rican Institute of Technology (ITCR), through the grant (ALECOSTA: Alemania-Costa Rica - PhD, 2019 (57463328)).

Beyond this, the authors declare that no other funds, grants, or financial support were received during the preparation of this manuscript.

Availability of Data and Material. The interactive GEE app previously described is available at: https://wawasita.users.earthengine.app/view/bambooprototypethailand

The source code generated during this research with JavaScript syntax and the output maps in GeoTiff format can be obtained from the following online repository: https://tubcloud.tu-berlin.de/s/cxP77CQERaR5ZPL

Specific output components can be requested from the research team, within a reasonable scope, by reaching the corresponding author at: [email protected].

Beck HE, Zimmermann NE, McVicar TR, et al (2018) Present and future Köppen-Geiger climate classification maps at 1-km resolution. Scientific Data 5:180214. https://doi.org/10.1038/sdata.2018.214
CEN/TC 350 (2013) Sustainability of Construction Works - Environmental Product Declarations - Core Rules for the Product Category of Construction Products
Chen L, Zhang X, Tang C, Gan X (2017) Effects of Site Factors on the Clonal Growth of Phyllostachys bambusoides f. shouzhu Yi. Journal of Plant Science and Phytopathology 1:071–079. https://doi.org/10.29328/journal.jpsp.1001009
Cheng X, Shi P, Hui C, et al (2015) An optimal proportion of mixing broad-leaved forest for enhancing the effective productivity of moso bamboo. Ecology and Evolution 5:1576–1584. https://doi.org/10.1002/ece3.1446
Churkina G, Organschi A, Reyer CPO, et al (2020) Buildings as a global carbon sink. Nature Sustainability 3:269–276. https://doi.org/10.1038/s41893-019-0462-4
Correia M, Carlos G, Rocha S (2013) Vernacular heritage and earthen architecture. CRC Press, Boca Raton
Da Silva Sobrinho Junior A, De Aragao Junior JL, Marden Torres S, et al (2009) Bamboo PH and Absorption in Different Liquids. Proceedings of the 11th International Conference on Non-conventional Materials and Technologies (NOCMAT 2009), Bath, UK
Devi AS, Singh KS (2021) Carbon storage and sequestration potential in aboveground biomass of bamboos in North East India. Scientific Reports 11: https://doi.org/10.1038/s41598-020-80887-w
Dida JJV, Araza AB, Eduarte GT, et al (2021) Towards nationwide mapping of bamboo resources in the Philippines: testing the pixel-based and fractional cover approaches. International Journal of Remote Sensing 42:3380–3404. https://doi.org/10.1080/01431161.2020.1871099
Dieren D (2020) Wood to City. City to Wood. MSc Thesis. Technical University Berlin.
Environmental Coalition on Standards ECOS (2020) TOWARDS A GREENER EU CONSTRUCTION PRODUCTS POLICY 2
Farr TG, Rosen PA, Caro E, et al (2007) The Shuttle Radar Topography Mission. Reviews of Geophysics 45: https://doi.org/10.1029/2005rg000183
Fu J (2001) Chinese Moso Bamboo: Its importance. B A M B O O · The Magazine of The American Bamboo Society 22:
Government of Western Australia (2014) Growing bamboos. In: . https://www.agric.wa.gov.au/growing-bamboos. Accessed 8 Nov 2020
International Resource Panel IRP (2017) Assessing Global Resource Use: Assessing global resource use: A systems approach to resource efficiency and pollution reduction. United Nations Environment Programme, Nairobi
International Union for Conservation of Nature (IUCN) (2010) Integrated Advanced Remote Sensing GIS Study for Bamboo-based Livelihood Analysis and Rural Development Planning in Nhamatanda, Dondo districts of Sofala province, Mozambique. FAO; INBAR; IUCN
Jarasiunas G (2016) Assessment of the agricultural land under steep slope in Lithuania. Journal of Central European Agriculture 17:176–187. https://doi.org/10.5513/JCEA01/17.1.1688
Kauschen JS, El-Khatib A, Granby-Larsen M, Larsson N (2020) Byggeriets Materialepyramide. In: . https://www.materialepyramiden.dk/. Accessed 24 Nov 2021
Kelchner SA (2013) Higher level phylogenetic relationships within the bamboos (Poaceae: Bambusoideae) based on five plastid markers. Molecular Phylogenetics and Evolution 67:404–413. https://doi.org/10.1016/j.ympev.2013.02.005
Kleinhenz V, Midmore D (2001) Aspects of Bamboo Agronomy. Advances in Agronomy, Newark
Li M, Dong Y, Jiang H, et al (2016) Tracking bamboo dynamics in Zhejiang, China, using time-series of Landsat data from 1990 to 2014. International Journal of Remote Sensing 37:1714–1729. https://doi.org/10.1080/01431161.2016.1165885
Lopez y Royo C, Baptista MJ, Webb S, et al (2021) Catalysing Science-based Policy action on Sustainable Consumption and Production – The value-chain approach & its application to food, construction and textiles. UNEP, Nairobi
Meijaard E, Brooks TM, Carlson KM, et al (2020a) The environmental impacts of palm oil in context. Nature Plants 6:1418–1426. https://doi.org/10.1038/s4147702000813w
Nath A, Lal R, Das A (2015) Ethnopedology and soil quality of bamboo (Bambusa sp.) based agroforestry system. Science of The Total Environment 521-522C: https://doi.org/10.1016/j.scitotenv.2015.03.059
OECD (2018) Global Material Resources Outlook to 2060 Economic Drivers and Environmental Consequences HIGHLIGHTS. OECD Publishing, Paris
Qi S, Song B, Liu C, et al (2022) Bamboo Forest Mapping in China Using the Dense Landsat 8 Image Archive and Google Earth Engine. Remote Sensing 14:762. https://doi.org/10.3390/rs14030762
Ramyarangsi S (1985) Bamboo Research in Thailand. In: https://www.inbar.int/. https://www.inbar.int/wp-content/uploads/2020/05/1493102465.pdf. Accessed 24 Nov 2021
Rusch F, Ceolin G, Hillig E (2019) Morphology, density and dimensions of bamboo fibers: a bibliographical compilation. Pesquisa Agropecuária Tropical 49: https://doi.org/10.1590/1983-40632019v4955007
Salonen S, Vangsbo P (2019) The challenges and potential of circular procurements in public construction projects. World Circular Economy Forum, Helsinki
Schröder S (2021) Dendrocalamus asper - Rough Bamboo. In: Guadua Bamboo. https://www.guaduabamboo.com/blog/dendrocalamus-asper. Accessed 22 Nov 2021
Shen L, Yang J, Zhang R, et al (2019) The benefits and barriers for promoting bamboo as a green building material in China—An integrative analysis. Sustainability 11: https://doi.org/10.3390/su11092493
Shi P, Preisler HK, Quinn BK, et al (2020) Precipitation is the most crucial factor determining the distribution of moso bamboo in Mainland China. Global Ecology and Conservation 22:e00924. https://doi.org/10.1016/j.gecco.2020.e00924
Song X, Zhou G, Jiang H, et al (2011) Carbon sequestration by Chinese bamboo forests and their ecological benefits: assessment of potential, problems, and future challenges. Environmental Reviews 19:418–428. https://doi.org/10.1139/a11-015
Thai Meteorological Department, Ministry of Digital Economy and Society (2021) Thailand Annual Weather Summary, BE 2563. In: https://www.tmd.go.th/. https://www.tmd.go.th/programs/uploads/yearlySummary/%E0%B8%AA%E0%B8%A3%E0%B8%B8%E0%B8%9B%E0%B8%AA%E0%B8%A0%E0%B8%B2 %E0%B8%A7%E0%B8%B0%E0%B8%AD%E0%B8%B2%E0%B8%81%E0%B8% B2%E0%B8%A8%E0%B8%9B%E0%B8%B5%202563.pdf. Accessed 24 Nov 2021
Udomsri S, Hoontrakul K (1997) Surface Texture of the Soil Series in Thailand. Academic papers, Soil Survey and Classification Group 412:
United Nations Environment Programme (2021) Catalysing Science-based Policy action on Sustainable Consumption and Production – The value-chain approach & its application to food, construction and textiles. UNEP
United States Department of Agriculture (1987) USDA Soil Textural Classification Study Guide. USDA Soil Conservation Service, Washington DC
van der Lugt P, Vogtlander J (2015) The environmental impact of industrial bamboo products - life-cycle assessment and carbon sequestration. INBAR Technical Report 35: https://doi.org/10.13140/RG.2.2.20797.46560
Verkerk PJ, Hasewaga M, Cramm M, et al (2021) The role of forest products in the global bioeconomy: Enabling substitution by wood-based products and contributing to the Sustainable Development Goals, 1st edn. Food & Agriculture Org., Rome
Verspeek S, van der Burgh F (2017) LCA-rapport Moso massieve plaat, Density, Bamboo X-treme (EN15804). Agrodome, The Netherlands
Vogtländer JG, van der Velden NM, van der Lugt P (2014) Carbon sequestration in LCA, a proposal for a new approach based on the global carbon cycle; cases on wood and on bamboo. The International Journal of Life Cycle Assessment 19:13–23. https://doi.org/10.1007/s11367-013-0629-6
Yiping L, Henley G (2010) Biodiversity in Bamboo Forests: a policy perspective for long term sustainability. INBAR
Yiping L, Yanxia L, Buckingham K, et al (2010) Bamboo and Climate Change Mitigation. INBAR, Beijing
You S, Zheng Q, Lin Y, et al (2020) Specific bamboo forest extraction and long-term dynamics as revealed by landsat time series stacks and google earth engine. Remote Sensing 12: https://doi.org/10.3390/rs12183095
Zhao Y, Feng D, Jayaraman D, et al (2018) Bamboo mapping of Ethiopia, Kenya and Uganda for the year 2016 using multi-temporal Landsat imagery. International Journal of Applied Earth Observation and Geoinformation 66:116–125. https://doi.org/10.1016/j.jag.2017.11.008
Zheng YS, Hong W (1998) Management of Phyllostachys Pubescens Stand. Xiamen University Publishing House, Xiamen

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Exploring the Potential for Bamboo Agroforestry through a Value Chain Approach in Thailand, using Cloud-based Data Management on Google Earth Engine

Status:

Version 1

Abstract

Figures

1 Introduction

1.1 Bamboo and Sustainable Construction

1.2 Bamboo Agroforestry

1.3 The Value Chain Approach

1.4 Potential vs. Existing Bamboo Area

2 Methods

2.1 Scale: From Thailand to Global

2.2 From Bamboo to Other Crops

2.3 Datasets

2.4 Calculations

2.5 Validation

3 Results

3.1 Final Overlay Map

3.2 Data Validation

3.3 Online App

4 Discussion

4.1 Availability of High-resolution Satellite Imagery

4.2 Different Species with Different Thresholds

4.3 Uses for Policymaking

Declarations

References

Additional Declarations

Status:

Version 1