Article

# Annual multi-resolution detection of land cover conversion to oil palm in the Peruvian Amazon

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

## Abstract

Oil palm expansion is a major threat to forest conservation in the tropics. Oil palm can also be a sustainable economic alternative if incentives for expansion outside forests are set in place. Consistent methods to monitor the time and location of oil palm expansion and the area converted from different land covers are essential for the success of such incentives. We developed methods to detect and quantify annual land cover changes associated with oil palm expansion in the Peruvian Amazon between 2001 and 2010 at two spatial scales and for two production modes. At the coarse scale, comprising the whole Peruvian Amazon, we used MODIS data to detect forest conversion to large-scale, industrial oil palm plantations based on metrics characterizing temporal changes in vegetation greenness associated with the conversion. At the fine scale, we used data from the satellite sensors Landsat TM/ETM + and ALOS-PALSAR to map and quantify the area from different land covers converted into large and small-scale oil palm plantations annually, in a focus area near the city of Pucallpa. Estimates were obtained from the elaboration and further combination of maps representing oil palm plantations by ages in 2010 and non-oil palm land covers in each year between 2001 and 2010. Validation data were obtained in the field and from geospatial information from previous studies. At the coarse scale, MODIS detected deforestation in 73% of training events larger than 50 ha. Detected events added up to 95% of the training areas. Total area converted to oil palm annually was quantified visually by using data from Landsat TM/ETM + with 96.3% accuracy. At the fine scale, the combination of data from Landsat TM/ETM + and ALOS-PALSAR identified oil palm expansion in areas larger than 5 ha with 94% accuracy and the year of expansion with an uncertainty of ± 1.3 years. This work underscores the need for data from multiple satellite sensors for a comprehensive monitoring of oil palm expansion, considering needs for information not only on the area expanded but also the time of conversion and land cover transitions associated with large- and small-scale plantations.

## No full-text available

... In particular, remote sensing data acquired through various earth observation satellites provide immense opportunity to monitor land use/land cover (LULC) changes caused by plantation cultivation. However, current state of the art methods based on remote sensing data are limited in their temporal frequency and scalability due to various reasons, such as need for human intervention, use of very simple machine learning methods and are applicable only for small regions (Hansen et al., 2008(Hansen et al., , 2013Hoscilo et al., 2011;Dong et al., 2012;Li and Fox, 2012;Margono et al., 2012;Miettinen et al., 2012a,b;Ziegler et al., 2012;Gutiérrez-Vélez and DeFries, 2013). To this date, there is no existing framework that can provide plantation extent maps in an automated fashion at yearly scales for large regions. ...
... Some methods involve extensive human involvement as they use visual interpretation in the detection process (Miettinen et al., 2012a,b;Ziegler et al., 2012). Few automatic machine learning based methods have also been proposed in the literature but they use very simple techniques such thresholding (Dong et al., 2012;Gutiérrez-Vélez and DeFries, 2013), nearest neighbor method (Li and Fox, 2012). Some sophisticated machine learning methods have only shown success in selected small-scale test dataset (Jia et al., 2017a,b). ...
Article
Full-text available
Plantation mapping is important for understanding deforestation and climate change. While most existing plantation products are created manually, in this paper we study an ensemble learning based framework for automatically mapping plantations in southern Kalimantan on a yearly scale using remote sensing data. We study the effectiveness of several components in this framework, including class aggregation, data sampling, learning model selection and post-processing, by comparing with multiple baselines. In addition, we analyze the quality of our plantation mapping product by visual examination of high resolution images. We also compare our method to existing manually labeled plantation datasets and show that our method can achieve a better balance of precision (i.e., user's accuracy) and recall (i.e., producer's accuracy).
... If the constant deforestation is not put to an end, the carbon content in our atmosphere would continue to rise breaking the near 420 ppm threshold further increasing the elevated events following climate change [20,6]. Remote sensing and satellite technologies can also help to determine where oil palm plantations can be effectively utilized as a sustainable economic alternative to other plant-based oils if appropriate incentives for expansion outside of forested areas area established [21]. To understand these issues, researchers have suggested that more focus needs to be given to the development of methods of monitoring and analysis of annual land cover changes associated with oil palm expansion [21]. ...
... Remote sensing and satellite technologies can also help to determine where oil palm plantations can be effectively utilized as a sustainable economic alternative to other plant-based oils if appropriate incentives for expansion outside of forested areas area established [21]. To understand these issues, researchers have suggested that more focus needs to be given to the development of methods of monitoring and analysis of annual land cover changes associated with oil palm expansion [21]. The use of highresolution imagery has also been used in conjunction with other techniques to further refine land use change in certain areas. ...
Conference Paper
Full-text available
Primary forests are continuously threatened day to day by urbanisation and the conversion to agricultural plantation such as palm oil production and other land uses being one of the major reasons. In SouthEast Asia, the widespread of oil palm has boomed over the last two decades, resulting in the downfall of tropical forest land. This change has been particularly prevalent in Borneo with protected lands increasingly developed for palm oil and already deforested lands are being converted into industrial plantations. The primary concerns relating to this pattern of land use change are the short and long-term impacts of logging on our natural environments and ecosystems and how patterns of deforestation are contributing to global environment issues such as climate change. By detecting and mapping logging activities, forestry departments are better able to predict land cover changes in a particular region as a result of development and deforestation. In this study, land cover assessment based on remote sensing techniques was used to analyse changes in the Bintulu district, Borneo especially oil palm growth and its influence on the decline of forest areas between 2016 and 2018. High resolution satellite imageries (3 m spatial resolution) from PlanetScope were used due to the benefits in helping to differentiate several land cover classes over a higher spatial resolution. Results showed that a decline of primary forests in Bintulu is about 26.5% over the past 2 years. With that decline, comes a rise of oil palm growth by 17.6% just within 2 years. An increased in logged areas (36.1%) for conversion to other land-covers with a steady decline of other land cover classes by at least 20% each year was detected. The accuracy of results proved a reasonable accuracy with 90.0% confidence with satellite imagery. Thus, with the result a necessary step can be considered to monitor and prevent deforestation and various encroachment in the forest. With high-resolution satellite data, monitoring at local scales has become possible to resource managers as a way to create timely and reliable assessments.
... Monitoring citrus planting dynamics has been difficult due to its high heterogeneity over space and through time (Gutiérrez-Vélez & DeFries, 2013;Vieira et al., 2012). Typically, national/regional statistics are the common data source for understanding citrus planting (Abdul et al., 2018). ...
... With the open-access of almost 50-year earth observation, Landsat has become the most suitable data source for monitoring land cover changes across multiple spatiotemporal scales Zhu et al., 2019). A few studies have been carried out to capture cash crop planting dynamics via multi-temporal Landsat data (Chen et al., 2016;Gutiérrez-Vélez and DeFries, 2013;Sun et al., 2017;Xiao et al., 2019;Xu et al., 2018b). More recently, Xu et al. (2018a) employed Landsat TM, ETM+, and OLI data to study the expansion of citrus planting at the annual scale from 1990 to 2016. ...
Article
Full-text available
Across Southeastern China, the expansion of citrus plantation has underpinned substantial economic development at the expense of forest cover and associated ecosystem services. The high-order complexity of citrus planting dynamics requires more detailed observations. To achieve this goal, we developed a novel monitoring scheme by integrating multi-epoch land cover classification and continuous change detection on the Google Earth Engine (GEE) platform using all available Landsat archives (1986–2018). First, a stable/changed area masking approach was adopted. Then, the GEE-based Continuous Change Detection and Classification (GEE-CCDC) algorithm was applied to detect the timing and position of citrus planting-related land cover changes. Finally, the effectiveness of the scheme was validated both spatially and temporally. We applied the proposed scheme in Xunwu, a typical County in Southeastern China where continuous expansion was disrupted by an outbreak of Huanglongbing (HLB, a destructive citrus disease). The validation results indicate that the multiepoch classification effectively identified the stable, abandoned, and newly cultivated areas related to citrus planting. Within these areas, our monitoring of citrus planting dynamics had an overall temporal accuracy of 90.59% and a mean detection date deviation of -13.21 days (leading). Based on the change detection results, we found a two-stage change pattern, namely a continuous expansion period (1986–2016) followed by a disturbance period (2017–2018). Moreover, almost half of the orchards were cultivated in regions with an elevation and slope of 301–400 m and 11–20◦. During the continuous expansion period, a total of 497.85 km2 natural lands were converted for citrus planting. Affected by the HLB outbreak, both orchard abandonment (114.66 km2) and new orchard cultivation (36.87 km2) were detected during the disturbance period. The proposed scheme offers a useful tool for cash crops management, and highlights the value of dense satellite data in monitoring the extent of ongoing human appropriation of natural ecosystems at a large scale and long-term.
... As an alternative, automated machine learning-based methods have been applied successfully to map plantation agriculture in the humid tropics. For example, Gutierrez Velez et al. [33] used a combination of MODIS Enhanced Vegetation Index (EVI) time-series and Landsat to detect forest conversion to oil palm in Peru. Automated approaches applied to detect tree crops have commonly utilized thresholding-based approaches [33][34][35], and the nearest neighbor method [36]. ...
... For example, Gutierrez Velez et al. [33] used a combination of MODIS Enhanced Vegetation Index (EVI) time-series and Landsat to detect forest conversion to oil palm in Peru. Automated approaches applied to detect tree crops have commonly utilized thresholding-based approaches [33][34][35], and the nearest neighbor method [36]. These approaches are limited in their capabilities, however, because they do not fully exploit crop-specific changes in vegetation characteristics over time (e.g., land clearing, growth cycles, and phenology specific to each tree crop species). ...
Article
Full-text available
Expansion of large-scale tree plantations for commodity crop and timber production is a leading cause of tropical deforestation. While automated detection of plantations across large spatial scales and with high temporal resolution is critical to inform policies to reduce deforestation, such mapping is technically challenging. Thus, most available plantation maps rely on visual inspection of imagery, and many of them are limited to small areas for specific years. Here, we present an automated approach, which we call Plantation Analysis by Learning from Multiple Classes (PALM), for mapping plantations on an annual basis using satellite remote sensing data. Due to the heterogeneity of land cover classes, PALM utilizes ensemble learning to simultaneously incorporate training samples from multiple land cover classes over different years. After the ensemble learning, we further improve the performance by post-processing using a Hidden Markov Model. We implement the proposed automated approach using MODIS data in Sumatra and Indonesian Borneo (Kalimantan). To validate the classification, we compare plantations detected using our approach with existing datasets developed through visual interpretation. Based on random sampling and comparison with high-resolution images, the user’s accuracy and producer’s accuracy of our generated map are around 85% and 80% in our study region.
... Since the mid-2000s, palm oil has become a growing threat to Amazonian forests, especially in Colombia, Ecuador, Peru, and the eastern part of the Brazilian Amazon (Furumo and Aide, 2017). Although palm oil plantations often replace other agricultural land uses, especially cattle ranching, it has been documented directly replacing primary forests (Castiblanco et al. 2013;de Almeida et al. 2020;Gutiérrez-Vélez and DeFries 2013). For example, between 2007 and 2013, 11% of deforestation in the Peruvian Amazon was driven by oil palm plantations (Vijay et al. 2018). ...
Chapter
This Report provides a comprehensive, objective, open, transparent, systematic, and rigorous scientific assessment of the state of the Amazon’s ecosystems, current trends, and their implications for the long-term well-being of the region, as well as opportunities and policy relevant options for conservation and sustainable development.
... Radar indices such as the polarization ratio, normalized difference index (NDI), and the NL index have also been used in forest type mapping [20,25,[27][28][29][30][31][32][33]. Another comprehensive approach that recent studies have frequently demonstrated is the combination of optical and SAR imagery [12,[19][20][21]27,28,34,35]. The synergy of structural information from SAR images and biophysical information from optical images has improved the accuracy and map detail. ...
Article
Full-text available
Highly detailed and accurate forest maps are important for various applications including forest monitoring, forestry policy, climate change, and biodiversity loss. This study demonstrates a comprehensive and geographically transferable approach to produce a 12 category high-resolution land use/land cover (LULC) map over mainland Vietnam in 2016 by remote sensing data. The map included several natural forest categories (evergreen broadleaf, deciduous (mostly deciduous broadleaf), and coniferous (mostly evergreen coniferous)) and one category representing all popular plantation forests in Vietnam such as acacia (Acacia mangium, Acacia auriculiformis, Acacia hybrid), eucalyptus (Eucalyptus globulus), rubber (Hevea brasiliensis), and others. The approach combined the advantages of various sensor data by integrating their posterior probabilities resulting from applying a probabilistic classifier (comprised of kernel density estimation and Bayesian inference) to each datum individually. By using different synthetic aperture radar (SAR) images (PALSAR-2/ScanSAR, PALSAR-2 mosaic, Sentinel-1), optical images (Sentinel-2, Landsat-8) and topography data (AW3D30), the resultant map achieved 85.6% for the overall accuracy. The major forest classes including evergreen broadleaf forests and plantation forests had a user's accuracy and producer's accuracy ranging from 86.0% to 95.3%. Our map identified 9.55 × 10 6 ha (±0.16 × 10 6 ha) of natural forests and 3.89 × 10 6 ha (±0.11 × 10 6 ha) of plantation forests over mainland Vietnam, which were close to the Vietnamese government's statistics (with differences of less than 8%). This study's result provides a reliable input/reference to support forestry policy and land sciences in Vietnam.
... More than 80 % of the global palm oil is produced in Indonesia and Malaysia (FAO, 2020). The governments of various developing countries in Asia, Africa and South America are promoting oil palm cultivation for energy, food security and poverty alleviation (Gutiérrez-Vélez and DeFries, 2013;Villela et al., 2014;Tarigan et al., 2015;Pirker et al., 2016). An area of 1.93 mha has been identified having potential for oil palm cultivation in India (Rethinam et al., 2012). ...
Article
The knowledge pertaining to phyto-availability of potassium (K), calcium (Ca) and magnesium (Mg) in soils and K, Ca and Mg concentrations in plant tissues and their relations in different productions systems of the world is limited. Understanding soil and leaf K, Ca and Mg concentrations and their stoichiometry in oil palm (Elaeis guineensis Jacq.), a globally demanding crop for vegetable oil production, is important for efficient K, Ca and Mg management. We, therefore, collected soil and leaf samples from oil palm plantations (OPP) established on three different soil series of India, and analysed to assess soil properties (pH, electrical conductivity, soil organic carbon) and concentrations of exchangeable K (Ex. K), Ca (Ex. Ca) and Mg (Ex. Mg) in soil and leaf K, Ca and Mg concentrations and their stoichiometry, and to visualize their relations. Soil properties, concentrations of Ex. K, Ex. Ca and Ex. Mg and their ratios in OPP of different soil series varied widely and differed significantly. The mean concentrations of Ex. K, Ex. Ca and Ex. Mg in OPP of Jangareddigudem and Madukkur series followed the order: Ex. Ca > Ex. Mg > Ex. K. However, soils of OPP of Karmali series had higher mean concentration of Ex. Ca followed by Ex. K followed by Ex. Mg. Leaf K, Ca and Mg concentration and their ratios in oil palm plantations of all the three-soil series varied widely and differed significantly. The mean concentrations of leaf K, Ca and Mg followed the order: Ca > K > Mg. There were positive and significant correlations of Ex. K with Ex. Ca in soils of Jangareddigudem and Madukkur series. Leaf Ca concentration was negatively and significantly correlated with leaf K concentration and Ex. K and Ex. Mg in soils of Jangareddigudem series. There were no correlations of leaf K, Ca and Mg concentration with Ex. K, Ex. Ca and Ex. Mg in soils of Madukkur and Karmali soil series. This study highlighted the need for maintaining equilibrium concentrations of Ex. K, Ex. Ca and Ex. Mg in soils for ensuring proper K, Ca and Mg nutrition of oil palm.
... Indeed, the spectral confusion between tree-based land cover types has been widely reported (Fagan et al., 2013(Fagan et al., , 2015Gutiérrez-Vélez and De Fries, 2013;Senf et al., 2013;Haro-Carrión and Southworth, 2018). As such, when teak plantations and forests were merged, and the image reclassified, the accuracy of the classification increased from 68 to 79 %, while the kappa coefficient increased from 0.60 to 0.70. ...
Article
Full-text available
Soil is a fundamental natural resource that is vital to the sustainable development of human societies. However, in many developing countries, increased intensity of use and inadequate land use planning has put a lot of pressure on marginal soil, leading to various forms of land degradation. The purpose of this study is to generate an integrated the land cover and terrain classification of the Ban Dan Na Kham watershed of Northern Thailand as a tool for sustainable land use planning. The watershed boundary and slope classes were delineated using the Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM). The slope was subsequently classified into gentle (<8 o), moderate (8-30 o) and steep (>30 o). The land cover map was generated through the supervised classification of Sentinel2 satellite imagery. Both map products were then integrated to provide the basis for land allocation and land use planning. The results show that 58 % of land currently under arable farming is either marginally suitable or practically unsuitable for that purpose. This ultimately leads to increased land degradation and soil loss. The land should consequently be reforested. Nevertheless, up to 10 km 2 of the watershed that is dedicated to other land use types-almost twice the current arable land area-is suitable for arable cropping. As such, given the proposed reforestation of the marginal and unsuitable arable lands, a large proportion of suitable land is still available to make up for the deficit. This will ultimately lead to increased productivity and reduced land degradation.
... Mapping oil palm distribution relying on optical satellite RS and in particular, the image-based approach has been reported as a challenging task due to the rapid oil palm canopy growth and spectral similarity of this plantation to the other land covers such as natural forest and rubber plantation Torbick et al. 2016;Víctor and Defries 2013). Relying on the temporal characteristics of optical sensors (phenology-based methods) also introduces some challenges particularly in small fragmented oil palm plantations as it usually uses Moderate Resolution Imaging Spectroradiometer (MODIS) data due to its high temporal resolution . ...
Article
Full-text available
The most crucial technical challenge facing the Malaysian and Indonesian oil palm industry is that the actual yield in the form of Fresh Fruit Bunch (FFB; unit in tonne per hectare (t ha −1)) are well below of potential levels and have stagnated over the last two decades. Closing this wide yield gap would have a positive impact on the revenue as it increases productivity per hectare and it eventually leads to less pressure on opening new land and mitigates environmental costs of production. With respect to the indispensable need for closing this gap for the future prosperity of this industry and sustainable production of palm oil, this study assessed oil palm yield, considering the potential growth of oil palm dependent on the site qualities and actual yield. Firstly, we mapped oil palm plantations combining yearly Advanced Land Observation Satellite (ALOS) Phased Array type L-band Synthetic Aperture Radar (PALSAR) and ALOS-2 mosaics of L-band backscatter, Moderate Resolution Imaging Spectroradiometer (MODIS) reflectance (MOD13Q1), and the MODIS Vegetation Continuous Field canopy cover product (MOD44B); where 10.3 and 6.68 million ha (Mha) of oil palm plantations were mapped, respectively, in Indonesia and Malaysia in 2017. Secondly, the age after planting was estimated at detected plantations using time series of MODIS canopy cover with correlation coefficient (r) of 0.68 and Root Mean Square Error (RMSE) of 4.7 years. Thirdly, the biophysical suitability of detected plantations was evaluated considering the spatial-temporal variation of different biophysical criteria. Combining information from second and third steps, we estimated the potential yield at 250 m spatial resolution. The average potential yield in Malaysia ranges between 13.8 t ha −1 and 19.3 t ha −1 in 2017, where in Indonesia it ranges between 17.8 t ha −1 and 21.7 t ha −1 in the same year. The actual yield in next step, has been quantified by HH-HV attribute of ALOS PALSAR and ALOS-2 mosaics, where the average actual yield in Malaysia ranges between 14.48 t ha−1 and 20.63 t ha −1 and in Indonesia it ranges between 8.49 t ha −1 and15.40 t ha −1 in 2017. Finally, comparing estimated potential and actual yields, we evaluated oil palm industries' performances where distinct differences were found between two countries. In most of the Malaysian states quantified actual yields were above or at the level of estimated potential yields, whereas in all Indonesian provinces quantified actual yields were well below the potential level. Considering the favourability of environment, among all provinces/states, Sabah, and Sarawak states in Malaysia and Aceh and North Kalimantan provinces in Indonesia ARTICLE HISTORY distinctly differ due to their poor performances from rest of provinces/ states. The information on different yields provided in this study are indispensable needs for efficient and accountable policies as it enables governors to directly target specific objectives such as subsidies on fertilizers, productive cultivars, and new technologies for the plantations suffering from low yield. Also, this study provides benchmarks for each province/state for scopes of actual yield improvements for long-term planning.
... Peru contains the second largest area of suitable land for oil palm production in the Amazon (Gutiérrez-Vélez et al., 2011). Within the Ucayali region of Peru, 12,100 ha was converted to oil palm plantations between 2001(Gutiérrez-Vélez & DeFries, 2013. Similarly, oil palm production has been growing in sub-Saharan Africa, expanding into 1.2 Mha between 1990 and 2017 (Ordway, Naylor, et al., 2017). ...
Article
Full-text available
Tropical forests account for a large portion of the Earth's terrestrial carbon pool. However, rapid deforestation threatens the stability of this carbon. We examine radiocarbon (Δ¹⁴C) and stable carbon (δ¹³C) isotopes of soil organic matter to provide insight into rates of carbon turnover, inputs, and losses of pasture‐derived (C4) versus forest or oil palm‐derived (C3) carbon. Data are presented for natural lowland forests on mineral soil converted to pastures in Peru and to oil palm plantations in Peru, Indonesia, and Cameroon. We additionally examine plots of secondary forests following agricultural use. There were large losses in carbon stocks under both pasture and oil palms. In the plots converted to pasture, our data indicate a preferential loss of relatively young carbon, and a greater loss of forest‐derived carbon than replacement with pasture‐derived carbon. Natural forests converted directly to oil palm plantations sustained losses in carbon, but Δ¹⁴C values suggest that the soil may retain a sufficient amount of newly acquired carbon to offset initial losses of young carbon. Furthermore, replacement of pastures with oil palm plantations facilitates the accumulation of young carbon, which may lead to a gradual increase in carbon stocks. The sites examined here are representative of the biophysical characteristics in roughly half of the humid tropics, suggesting that these findings may be applicable to a large area of similarly managed mineral soils in lowland tropical forests.
... Each decision tree produces a result for each pixel that is considered as a vote. The final prediction for each pixel corresponds to the most voted class among all the decision trees [28,46]. For this application, we used the Random Forests package [47] implemented in the R computing environment. ...
Article
Full-text available
Tropical forests are disappearing at unprecedented rates, but the drivers behind this transformation are not always clear. This limits the decision-making processes and the effectiveness of forest management policies. In this paper, we address the extent and drivers of deforestation of the Choco biodiversity hotspot, which has not received much scientific attention despite its high levels of plant diversity and endemism. The climate is characterized by persistent cloud cover which is a challenge for land cover mapping from optical satellite imagery. By using Google Earth Engine to select pixels with minimal cloud content and applying a random forest classifier to Landsat and Sentinel data, we produced a wall-to-wall land cover map, enabling a diagnosis of the status and drivers of forest loss in the region. Analyses of these new maps together with information from illicit crops and alluvial mining uncovered the pressure over intact forests. According to Global Forest Change (GFC) data, 2324 km2 were deforested in this area from 2001 to 2018, reaching a maximum in 2016 and 2017. We found that 68% of the area is covered by broadleaf forests (67,473 km2) and 15% by shrublands (14,483 km2), the latter with enormous potential to promote restoration projects. This paper provides a new insight into the conservation of this exceptional forest with a discussion of the drivers of forest loss, where illicit crops and alluvial mining were found to be responsible for 60% of forest loss.
... In the Amazon region, one of the most mega-diverse regions on earth (Foley et al. 2007), land use changes have been linked to neoliberal agrarian policies (Arce-Nazario 2007), smallholder resettlement schemes and policy incentives (Bennett et al. 2018), road and infrastructure development (Andersen et al. 2002;Perz et al. 2013), cattle farming and cultivation of illicit crops (Armenteras et al. 2006). Policy and push-pull factors, including access to land, road development and coca production, but also violence and poverty in their former circumstances attracted colonist settlers, called mestizo households (Alvarez and Naughton-Treves 2003;Chavez and Perz 2012;Guevara Salas 2009;Labarta et al. 2008). Cropland expansion has also been identified as a major driver of land cover change (Gutiérrez-Vélez and DeFries 2013), providing high returns especially in areas of previously intact forests (Butler and Laurance 2009), but also threatening ecosystem services (Srinivas and Koh 2016). ...
Article
Full-text available
Few longitudinal studies link agricultural biodiversity, land use and food access in rural landscapes. In this paper, we test the hypothesis that, in a context of economic change, cash crop expansion is associated with deforestation, reduced agrobiodiversity and changes in food access. For this purpose, we analysed data collected from the same 53 upland and floodplain mestizo households in Ucayali, Peru, in 2000 and 2015. We found an emerging transition towards less diversified food access coupled with loss of forest cover and reduced agricultural biodiversity. In 2015, diets appeared to rely on fewer food groups, fewer food items, and on products increasingly purchased in the market compared to 2000. Wild fruits and plants were mentioned, but rarely consumed. Agricultural production systems became more specialised with a shift towards commercial crops. Peak deforestation years in the 15-year period appeared linked with incentives for agricultural expansion. Our results suggest an overall trend from diversified productive and “extractive” systems and more diverse food access, towards specialized productive systems, with less diverse food access and stronger market orientation (both in production and consumption). The assumption in the food and agricultural sciences that increased income and market-orientation is linked to improved food security, is challenged by our integrated analyses of food access, agrobiodiversity, land use and forest cover. Our results highlight the importance of longitudinal, multidimensional, systemic analyses, with major implications for land use, food and health policies. The potential risks of parallel homogenisation of diets and agricultural production systems require interdisciplinary research and policies that promote integrated landscape approaches for sustainable and inclusive food systems.
... With its greater per-hectare production and economic competitiveness, oil palm is a pivotal crop in ensuring sufficient edible oil is available in the global market [1]. Agriculture has at times become a controversial topic among conservationists due to its negative impacts on the environment, such as biodiversity loss, deforestation, and increased carbon emissions [2][3][4][5]. Precision agriculture (PA), which involves informed decision making in agriculture using information interpreted from sensor-based data (such as the remote sensing data in this paper) or other sources, is currently sought as a solution for improved sustainable food production [6][7][8][9]. 2 of 28 In fertilizer application, PA enables site specific management by determining macronutrient status and fertilizer requirement in individual plants. ...
Article
Full-text available
Oil palm crops are essential for ensuring sustainable edible oil production, in which production is highly dependent on fertilizer applications. Using Landsat-8 imageries, the feasibility of macronutrient level classification with Machine Learning (ML) was studied. Variable rates of compost and inorganic fertilizer were applied to experimental plots and the following nutrients were studied: nitrogen (N), phosphorus (P), potassium (K), magnesium (Mg) and calcium (Ca). By applying image filters, separability metrics, vegetation indices (VI) and feature selection, spectral features for each plot were acquired and used with ML models to classify macronutrient levels of palm stands from chemical foliar analysis of their 17th frond. The models were calibrated and validated with 30 repetitions, with the best mean overall accuracy reported for N and K at 79.7 ± 4.3% and 76.6 ± 4.1% respectively, while accuracies for P, Mg and Ca could not be accurately classified due to the limitations of the dataset used. The study highlighted the effectiveness of separability metrics in quantifying class separability, the importance of indices for N and K level classification, and the effects of filter and feature selection on model performance, as well as concluding RF or SVM models for excessive N and K level detection. Future improvements should focus on further model validation and the use of higher-resolution imaging.
... These approaches are, however, very labor intensive, which limits their utility for upscaling and developing a near realtime monitoring system. Some studies have employed radar imagery, which have benefitted from penetration through clouds, improved resolution and high revisiting frequency [23][24][25] . Recent experimental studies have demonstrated the usability of free and open Sentinel data to detect oil palm plantations 26 . ...
Preprint
Full-text available
In recent decades, global oil palm production has shown an abrupt increase, with almost 90% produced in Southeast Asia alone. Monitoring oil palm is largely based on national surveys and inventories or one-off mapping studies. However, they do not provide detailed spatial extent or timely updates and trends in oil palm expansion or age. Palm oil yields vary significantly with plantation age, which is critical for landscape-level planning. Here we show the extent and age of oil palm plantations for the year 2017 across Southeast Asia using remote sensing. Satellites reveal a total of 11.66 (+/- 2.10) million hectares (Mha) of plantations with more than 45% located in Sumatra. Plantation age varies from ~7 years in Kalimantan to ~13 in Insular Malaysia. More than half the plantations on Kalimantan are young (<7 years) and not yet in full production compared to Insular Malaysia where 45% of plantations are older than 15 years, with declining yields. For the first time, these results provide a consistent, independent, and transparent record of oil palm plantation extent and age structure, which are complementary to national statistics.
... Vast amount of the geospatial remote sensing data provided in the GEE has allowed the powerful cloud-based platform to be used in various studies involving deforestation, oil palm plantations, environmental assessment, change detection and urban classifications (Patel et al., 2015;Dong et al., 2016;Goldblatt et al., 2016;Shelestov et al., 2017). GEE can be accessed either through Application Programming Interface (API) or web-based Interactive Development Environment (IDE) (Gorelick et al., 2017). The data catalog provided in the GEE houses a multi-petabyte accessible geospatial dataset that is made up of Earth-observing remote sensing images, including Landsat, MODIS, Sentinel-1 and Sentinel-2. ...
Article
Oil palm plays a pivotal role in the ecosystem, environment, economy and without proper monitoring, uncontrolled oil palm activities could contribute to deforestation that can cause high negative impacts on the environment and therefore, proper management and monitoring of the oil palm industry are necessary. Mapping the distribution of oil palm is crucial in order to manage and plan the sustainable operations of oil palm plantations. Remote sensing provides a means to detect and map oil palm from space effectively. Recent advances in cloud computing and big data allow rapid mapping to be performed over large a geographical scale. In this study, 30 m Landsat 8 data were processed using a cloud computing platform of Google Earth Engine (GEE) in order to classify oil palm land cover using non-parametric machine learning algorithms such as Support Vector Machine (SVM), Classification and Regression Tree (CART) and Random Forest (RF) for the first time over Peninsular Malaysia. The hyperparameters were tuned, and the overall accuracy produced by the SVM, CART and RF were 93.16%, 80.08% and 86.50% respectively. Overall, the SVM classified the 7 classes (water, built-up, bare soil, forest, oil palm, other vegetation and paddy) the best. However, RF extracted oil palm information better than the SVM. The algorithms were compared and the McNemar's test showed significant values for comparisons between SVM and CART and RF and CART. On the other hand, the performance of SVM and RF are considered equally effective. Despite the challenges in implementing machine learning optimisation using GEE over a large area, this paper shows the efficiency of GEE as a cloud-based free platform to perform bioresource distributions mapping such as oil palm over a large area in Peninsular Malaysia.
... Palm oil production has historically been geographically concentrated in Indonesia and Malaysia, while consumption is distributed globally (Figures 1 and 2). With growing demand, suitable land in the leading producer countries is becoming scarce [88], and plantings have expanded in Africa and Latin America [89][90][91][92], often promoted by national governments [93]. While evidence on direct livelihoods impacts are limited, some recent studies find that smallholder oil palm production improves both village-level assets [94] and household-level livelihoods and nutrition, though the effects are stronger for more affluent households and are driven in large part by farm expansion [95]. ...
Article
Full-text available
Multi-stakeholder initiatives (MSIs) are a form of private governance sometimes used to manage the social and environmental impacts of supply chains. We argue that there is a potential tension between input and output legitimacy in MSIs. Input legitimacy requires facilitating representation from a wide range of organizations with heterogeneous interests. This work, however, faces collective action problems that could lead to limited ambitions, lowering output legitimacy. We find that, under the right conditions a relatively small group of motivated actors, who we call institutional stewards, may be willing to undertake the cost and labor of building and maintaining the MSI. This can help reconcile the tension between input and output legitimacy in a formal sense, though it also results in inequalities in power. We test this claim using a case study of organizations’ activities in the Roundtable on Sustainable Palm Oil (RSPO). We find that a small group of founding members—and other members of long tenure—account for a disproportionate level of activity in the organization.
... Anderson et al (2018) states that due to lack of coordination, the distinct government agencies in Peru allocate the same land for different uses but this is primarily assumed to be unintentional, and even if intentional, it is presumed to cause no conflict [49]. Gutierrez-Velez et al (2011) observed that companies that intend to expand their palm oil plantations in Peru, seek to do so in areas that have not been assigned any other use to avoid conflict [50]. Therefore, we believe that the lower rates of forest loss in cases of overlapping land allocations in this region could also be a conflict avoidance strategy between land users, and that these overlapping allocations could be used as a policy tool for conservation [49]. ...
Article
Full-text available
Over the past decades, the Peruvian Amazon has experienced a rapid change in forest cover due to the expansion of agriculture and extractive activities. This study uses spectral mixture analysis (SMA) in a cloud-computing platform to map forest loss within and outside indigenous territories, protected areas, mining concessions, and reforestation concessions within the Madre de Dios Region in Peru. The study area is focused on key areas of forest loss in the western part of the Tambopata National Reserve and surrounding the Malinowski River. Landsat 8 Operational Land Imager and Landsat 7 Enhanced Thematic Mapper Plus surface reflectance data spanning 2013–2018 were analyzed using cloud-based SMA to identify patterns of forest loss for each year. High-resolution Planet Dove (3m) and RapidEye (5m) imagery were used to validate the forest loss map and to identify the potential drivers of loss. Results show large areas of forest loss, especially within buffer zones of protected areas. Forest loss also appears in the Kotsimba Native Community within a 1 km buffer of the Malinowski River. In addition to gold mining, agriculture and pasture fields also appear to be major drivers of forest loss for our study period. This study also suggests that gold mining activity is potentially not restricted to the legal mining concession areas, with 49% of forest loss occurring outside the mining concessions. Overall accuracy obtained for the forest loss analysis was 96%. These results illustrate the applicability of a cloud-based platform not only for land use land cover change detection but also for accessing and processing large datasets; the importance of monitoring not only forest loss progression in the Madre de Dios, which has been increasing over the years, especially within buffer zones, but also its drivers; and reiterates the use of SMA as a reliable change detection classification approach.
... Further, Latin America has the largest remaining forested area in the world suitable for oil palm cultivation (Furumo and Aide, 2017). Some authors have suggested that conversion to oil palm may be among the major drivers of the next wave of habitat change in the Neotropical lowlands (Gutiérrez-Vélez and DeFries, 2013;Lees and Vieira, 2013). While the majority of studies on the impacts of oil palm plantations on biodiversity have taken place in Southeast Asia, recent work in Latin America shows similar patterns to those found in Asia in terms of major reductions in bird, mammal, amphibian and invertebrate species in oil palm compared to native forest (Gallmetzer and Schulze, 2015;Lees et al., 2015;Juen et al., 2016;Mendes-Oliveira et al., 2017;Pardo et al., 2018a). ...
Article
Full-text available
Oil palm (Elaeis guineensis) plantations are one of the most rapidly expanding agroecosystems in the tropics, including Latin America. While many studies have demonstrated that large oil palm monocultures (>100 ha) are detrimental to biodiversity, including mammals, little is known about the impact of small-scale oil palm plantations, especially in the Neotropics. Here we used a camera trapping survey to compare species richness, community structure, and relative abundances of mid to large-bodied terrestrial mammals in small-scale oil palm plantations (<100 ha) and secondary forest fragments within a highly modified landscape mosaic in the southeastern lowlands of Tabasco, Mexico. Contrary to our expectations, we found no differences in the overall mammal communities between the oil palm and forest fragments, including species richness or mean relative abundance. Individual species showed some apparent differences in their total detections between the two habitats, with 11 having greater detections in forest than oil palm, and only two with greater detections in oil palm. Further, oil palm sites were more similar to one another in terms of mammal community structure than the secondary forest fragments. We found that shorter distance to forest patches was related to higher mammal species richness in both forest fragments and oil palm plantations. Twelve terrestrial mammal species known to occur in forested areas in the state of Tabasco were never detected in either vegetation type in our surveys, highlighting the fact that the mammal community in this landscape had already been reduced to those species most resilient to human disturbance. Our findings suggest that small-scale oil palm plantations in this region are used at least to some degree by most mammals that are also found in the remaining secondary forest fragments in this landscape, but that access to nearby forest is important for these species. In order to recover more of the original mammal community of the region and prevent further reductions in biodiversity, conservation priorities should center around reducing hunting pressure, allowing forest regeneration and increasing connectivity between protected areas and along waterways.
... Classification of time series using Landsat Thematic Mapper (TM) and ALOS PALSAR was successful and the oil palm trees were able to be clustered into several groups of age via RF [50]. Study on basal stem rot disease detection in oil palm plantations via RF model achieved the highest accuracy and the best result compared to SVM and CART models [51]. ...
Article
Biomass is a promising resource in Malaysia for energy, fuels, and high value-added products. However, regards to biomass value chains, the numerous restrictions and challenges related to the economic and environmental features must be considered. The major concerns regarding the enlargement of biomass plantation is that it requires large amounts of land and environmental resources such as water and soil that arises the danger of creating severe damages to the ecosystem (e.g. deforestation, water pollution, soil depletion etc.). Regarded concerns can be diminished when all aspects associated with palm biomass conversion and utilization linked with environment, food, energy and water (EFEW) nexus to meet the standard requirement and to consider the potential impact on the nexus as a whole. Therefore, it is crucial to understand the detail interactions between all the components in the nexus once intended to look for the best solution to exploit the great potential of biomass. This paper offers an overview regarding the present potential biomass availability for energy production, technology readiness, feasibility study on the techno-economic analyses of the biomass utilization and the impact of this nexus on value chains. The agro-biomass resources potential and land suitability for different crops has been overviewed using satellite imageries and the outcomes of the nexus interactions should be incorporated in developmental policies on biomass. The paper finally discussed an insight of digitization of the agriculture industry as future strategy to modernize agriculture in Malaysia. Hence, this paper provides holistic overview of biomass competitiveness for sustainable bio-economy in Malaysia.
... More than 80 % of the global palm oil is produced in Indonesia and Malaysia (FAO, 2020). The governments of various developing countries in Asia, Africa and South America are promoting oil palm cultivation for energy, food security and poverty alleviation (Gutiérrez-Vélez and DeFries, 2013;Villela et al., 2014;Tarigan et al., 2015;Pirker et al., 2016). An area of 1.93 mha has been identified having potential for oil palm cultivation in India (Rethinam et al., 2012). ...
... The availability of radar imagery has given rise to a new generation of remote sensing studies benefiting from penetration of clouds, improved resolution and high revisiting frequency [18][19][20] . Recent experimental studies have demonstrated the usability of the European Union's free and open Copernicus Sentinel data to detect oil palm plantations [21][22][23] . ...
Article
Full-text available
In recent decades, global oil palm production has shown an abrupt increase, with almost 90% produced in Southeast Asia alone. To understand trends in oil palm plantation expansion and for landscape-level planning, accurate maps are needed. Although different oil palm maps have been produced using remote sensing in the past, here we use Sentinel 1 imagery to generate an oil palm plantation map for Indonesia, Malaysia and Thailand for the year 2017. In addition to location, the age of the oil palm plantation is critical for calculating yields. Here we have used a Landsat time series approach to determine the year in which the oil palm plantations are first detected, at which point they are 2 to 3 years of age. From this, the approximate age of the oil palm plantation in 2017 can be derived.
... Some use only Synthetic Aperture Radar (SAR) [18,11], which is unaffected by the frequent cloud coverage in tropical regions [10]. Perhaps even more popular is the combination of optical and Radar images [19,20,21,22,23,12,17,24]. Only a few works map oil palms at country-wide or continental scales. ...
Preprint
Full-text available
Accurate mapping of oil palm is important for understanding its past and future impact on the environment. We propose to map and count oil palms by estimating tree densities per pixel for large-scale analysis. This allows for fine-grained analysis, for example regarding different planting patterns. To that end, we propose a new, active deep learning method to estimate oil palm density at large scale from Sentinel-2 satellite images, and apply it to generate complete maps for Malaysia and Indonesia. What makes the regression of oil palm density challenging is the need for representative reference data that covers all relevant geographical conditions across a large territory. Specifically for density estimation, generating reference data involves counting individual trees. To keep the associated labelling effort low we propose an active learning (AL) approach that automatically chooses the most relevant samples to be labelled. Our method relies on estimates of the epistemic model uncertainty and of the diversity among samples, making it possible to retrieve an entire batch of relevant samples in a single iteration. Moreover, our algorithm has linear computational complexity and is easily parallelisable to cover large areas. We use our method to compute the first oil palm density map with $10\,$m Ground Sampling Distance (GSD) , for all of Indonesia and Malaysia and for two different years, 2017 and 2019. The maps have a mean absolute error of $\pm$7.3 trees/$ha$, estimated from an independent validation set. We also analyse density variations between different states within a country and compare them to official estimates. According to our estimates there are, in total, $>1.2$ billion oil palms in Indonesia covering $>$15 million $ha$, and $>0.5$ billion oil palms in Malaysia covering $>6$ million $ha$.
... Since the mid-2000s, palm oil has become a growing threat to Amazonian forests, especially in Colombia, Ecuador, Peru, and the eastern part of the Brazilian Amazon (Furumo and Aide, 2017). Although palm oil plantations often replace other agricultural land uses, especially cattle ranching, it has been documented directly replacing primary forests (Castiblanco et al. 2013;de Almeida et al. 2020;Gutiérrez-Vélez and DeFries 2013). For example, between 2007 and 2013, 11% of deforestation in the Peruvian Amazon was driven by oil palm plantations (Vijay et al. 2018). ...
Technical Report
Full-text available
Deforestation, the complete removal of an area’s forest cover; and forest degradation, the significant loss of forest structure, functions, and processes; are the result of the interaction between various direct drivers, often operating in tandem. By 2018, the Amazon biome had lost approximately 870,000 km2 of its original forest cover, mainly due to agricultural expansion. Other direct drivers of forest loss include the opening of new roads, construction of hydroelectric dams, exploitation of minerals and oil, and urbanization. Impacts of deforestation range from local to global, including local changes in landscape configuration, climate, and biodiversity; regional impacts on hydrological cycles; and global increase of greenhouse gas emissions. Of the remaining Amazonian forests, 17% are degraded, corresponding to approximately 1,036,080 km2. Various anthropogenic drivers, including understory fires, edge effects, selective logging, hunting, and climate change can cause forest degradation. Degraded forests have significantly different structure, microclimate, and biodiversity as compared to undisturbed ones. These forests tend to have higher tree mortality, lower carbon stocks, more canopy gaps, higher temperatures, lower humidity, higher wind exposure, and exhibit compositional and functional shifts in both fauna and flora. Degraded forests can come to resemble their undisturbed counterparts, but this depends on the type, duration, intensity, and frequency of the disturbance event. In some cases, this may prohibit the return to a historic baseline. Avoiding further loss and degradation of Amazonian forests is crucial to ensure they continue to provide valuable and life-supporting ecosystem services.
... However, as an optical sensor, Landsat data are severely influenced by the presence of clouds, particularly in the tropics 53,54 . Furthermore, the spectral resolution of single-date Landsat data has often not been sufficient by itself to separate forest and tree plantations 13,55 . Radar data, in contrast, are less sensitive to cloud cover and potentially more sensitive to differences in structure between natural and anthropogenic tree cover 56,57 . ...
Article
Full-text available
Across the tropics, recent agricultural shifts have led to a rapid expansion of tree plantations, often into intact forests and grasslands. However, this expansion is poorly characterized. Here, we report tropical tree plantation expansion between 2000 and 2012, based on classifying nearly 7 million unique patches of observed tree cover gain using optical and radar satellite imagery. The resulting map was a subsample of all tree cover gain but we coupled it with an extensive random accuracy assessment (n = 4,269 points) to provide unbiased estimates of expansion. Most predicted gain patches (69.2%) consisted of small patches of natural regrowth (31.6 ± 11.9 Mha). However, expansion of tree plantations also dominated increases in tree cover across the tropics (32.2 ± 9.4 Mha) with 92% of predicted plantation expansion occurring in biodiversity hotspots and 14% in arid biomes. We estimate that tree plantations expanded into 9.2% of accessible protected areas across the humid tropics, most frequently in southeast Asia, west Africa and Brazil. Given international tree planting commitments, it is critical to understand how future tree plantation expansion will affect remaining natural ecosystems.
... Since the mid-2000s, palm oil has become a growing threat to Amazonian forests, especially in Colombia, Ecuador, Peru, and the eastern part of the Brazilian Amazon (Furumo and Aide, 2017). Although palm oil plantations often replace other agricultural land uses, especially cattle ranching, it has been documented directly replacing primary forests (Castiblanco et al. 2013;de Almeida et al. 2020;Gutiérrez-Vélez and DeFries 2013). For example, between 2007 and 2013, 11% of deforestation in the Peruvian Amazon was driven by oil palm plantations (Vijay et al. 2018). ...
Chapter
Full-text available
This Report provides a comprehensive, objective, open, transparent, systematic, and rigorous scientific assessment of the state of the Amazon’s ecosystems, current trends, and their implications for the long-term well-being of the region, as well as opportunities and policy relevant options for conservation and sustainable development.
... Furthermore, according to Shaharum et al., 76 the RF algorithm derived oil palm information better than the other classifier used, and the RF algorithm enhanced classification when the RF trees were ensembled into a forest. 69 Moreover, according to Attarchi and Gloaguen, 124 nonparametric RF models can accommodate a large number of correlated input variables and biased data, while avoiding overfitting. As a result, RF models are highly useful for mapping heterogeneous landscapes. ...
Article
Full-text available
Oil palm phenology has many advantages in managing the sustainability of oil palm plantations. The phenology of oil palms is a key issue in harvest estimation, fruit bunch production, estimating oil palm taxes, replanting, fertilization, and detecting oil palm disease. One of the recently developed methods of oil palm phenology involves the use of remote sensing technology. We evaluated and reviewed the current state of oil palm phenology based on remote sensing and conducted an optimized systematic review of recent scientific publications, specifically focusing on scientific peer-reviewed papers published between 1990 and 2021, comprising over 100 existing journal papers on remote sensing for oil palm phenology. The review includes a description of the state of the art and the mapping of oil palm phenology based on sensors, biophysical tree parameters, and classification techniques and also describes the state of the art in the development of regression models of oil palm phenology based on wavelength, biophysical tree parameters, and the type of regression model. Finally, the review provided an opportunity to develop suitable techniques for the identification, classification, and the construction of regression models of oil palm phenology. There is a lot of potential in combining multisensor approaches, suitable classification methods, and regression models for oil palm phenology. For future studies on oil palm phenology, we recommend integrating machine learning with oil palm biophysical parameters based on multisensor remote sensing technologies.
... A combination of radar with optical imaging has shown promising results, as SAR can penetrate clouds and cirrus clouds (Carneiro et al., 2020;Hirschmugl et al., 2020;Rao et al., 2021). Platforms, such as Sentinel-1 (radar) and Sentinel-2 (optical), are a new data-rich source that can be used in LULC mapping to monitor crop phenology, as well as to detect cultivated and irrigated plots, with high spatial and temporal resolutions (Gutiérrez-Vélez & Defries, 2013;Kou et al., 2015;Tian et al., 2019). ...
Article
Full-text available
Land-use and land-cover (LULC) are important environmental properties of the Earth’s surface. Satellite platforms and state-of-the-art algorithms enable the mapping of large areas, but they still need to be improved. This study aimed to compare free- and open-access images from radar and optical sensors, using the Google Earth Engine™ (GEE) for supervised classification of LULC for five municipalities in Roraima State, Brazil. Sentinel-1 (S1) scenes were used along with Landsat 8 (L8) and Sentinel-2 (S2) ones, resulting in five classification approaches S1 (SD), L8 (ODL), S2 (ODS), S1+L8 (SODL), and S1+S2 (SODS), with an auxiliary ALOS World 3D dataset (DEM≈30m). Accuracy was assessed by an error matrix. The SD approach was significantly different (P ≤ 0.01) from the others using a mean F1-score of 0.80. ODL and ODS had barely perceptible differences (P ≤ 0.1), showing F1-scores of 0.95 and 0.92, respectively. When comparing ODL (F1=0.95) and SODL (F1=0.95) no difference was found (P > 0.1). However, by comparing ODS (F1=0.92) and SODS (F1=0.94), there was a significant classification improvement (P ≤ 0.05). In short, the approaches SODL and SODS had the best pixel-based supervised classification performance. KEYWORDS Machine learning; Sentinel-1; Sentinel-2; Landsat 8; Savanna
... Since the mid-2000s, palm oil has become a growing threat to Amazonian forests, especially in Colombia, Ecuador, Peru, and the eastern part of the Brazilian Amazon (Furumo and Aide, 2017). Although palm oil plantations often replace other agricultural land uses, especially cattle ranching, it has been documented directly replacing primary forests (Castiblanco et al. 2013;de Almeida et al. 2020;Gutiérrez-Vélez and DeFries 2013). For example, between 2007 and 2013, 11% of deforestation in the Peruvian Amazon was driven by oil palm plantations (Vijay et al. 2018). ...
Chapter
Full-text available
This chapter discusses the main drivers of deforestation and forest degradation in the Amazon, particularly agricultural expansion, road construction, mining, oil and gas development, forest fires, edge effects, logging, and hunting. It also examines these activities’ impacts and synergies between them.
... Since the mid-2000s, palm oil has become a growing threat to Amazonian forests, especially in Colombia, Ecuador, Peru, and the eastern part of the Brazilian Amazon (Furumo and Aide, 2017). Although palm oil plantations often replace other agricultural land uses, especially cattle ranching, it has been documented directly replacing primary forests (Castiblanco et al. 2013;de Almeida et al. 2020;Gutiérrez-Vélez and DeFries 2013). For example, between 2007 and 2013, 11% of deforestation in the Peruvian Amazon was driven by oil palm plantations (Vijay et al. 2018). ...
Book
Full-text available
The Science Panel for the Amazon (SPA) is an unprecedented initiative convened under the auspices of the United Nations Sustainable Development Solutions Network (SDSN). The SPA is composed of over 200 preeminent scientists and researchers from the eight Amazonian countries, French Guiana, and global partners. These experts came together to debate, analyze, and assemble the accumulated knowledge of the scientific community, Indigenous peoples, and other stakeholders that live and work in the Amazon. The Panel is inspired by the Leticia Pact for the Amazon. This is a first-of-its-kind Report which provides a comprehensive, objective, open, transparent, systematic, and rigorous scientific assessment of the state of the Amazon’s ecosystems, current trends, and their implications for the long-term well-being of the region, as well as opportunities and policy relevant options for conservation and sustainable development. The three volumes of the final report can be downloaded from: https://www.theamazonwewant.org/amazon-assessment-report-2021/
... Since the mid-2000s, palm oil has become a growing threat to Amazonian forests, especially in Colombia, Ecuador, Peru, and the eastern part of the Brazilian Amazon (Furumo and Aide, 2017). Although palm oil plantations often replace other agricultural land uses, especially cattle ranching, it has been documented directly replacing primary forests (Castiblanco et al. 2013;de Almeida et al. 2020;Gutiérrez-Vélez and DeFries 2013). For example, between 2007 and 2013, 11% of deforestation in the Peruvian Amazon was driven by oil palm plantations (Vijay et al. 2018). ...
Chapter
Full-text available
This chapter presents country-specific descriptions of human intervention in the Amazon. In general, a rapid expansion of agricultural and extractive activities, mostly for export but also for domestic markets, and to a lesser degree small scale agriculture, have led to extensive deforestation and environmental degradation without substantially improving the living conditions of the population. Government policies and the extent of State ascendancy in the area also seem to be a powerful determinant of the nature and scale of the process. Despite the common underlying international and domestic economic and political forces in the Amazon, each country has its own particularities. In the case of Colombia, the process was shaped by the guerilla presence and deteriorated after the Peace Treaty, which does not mention “deforestation” and perpetuates Colombia’s extractivist model. Ecuador’s case is representative of the link between fossil fuel extraction, environmental deterioration, and social exclusion. The case of Peru shows an Amazon perceived as a territory awaiting to be “conquered, occupied, and exploited”, subjected to an unwavering extractive and market-orientated drive. In Bolivia, contradictions between conservation and state-led development policies and business activities, which have transformed it into the second deforestation hotspot of Amazonia after Brazil, are presented. The Venezuelan Amazon is subject to rampant violence and illegal activity driven by the political geography of gold in mixed configurations of governance, with blurred boundaries between legality and illegality and prevailing negligence concerning conservation. The Guianas share low deforestation levels and lower environmental pressures, but the recent expansion of gold mining poses a serious threat. The Brazilian case presented in the previous Chapter is referenced here when comparing countries’ experienes. Conservation experiences are also included. In all cases, unsustainable extractivist models have outpaced conservation policies; however, these experiences can prove useful in the design of effective conservation policies, reduction of greenhouse gas emissions, and improvements in living conditions of Indigenous peoples and local communities.
... However, as an optical sensor, Landsat data are severely in uenced by the presence of clouds, particularly in the tropics (51,52). Furthermore, the spectral resolution of single-date Landsat data has often not been su cient by itself to separate forest and tree plantations (14,53). ...
Preprint
Full-text available
Across the tropics, recent agricultural shifts have led to a rapid expansion of tree plantations, often into intact forest and grassland habitats. However, this expansion is poorly characterized. Here we report tropical tree plantation expansion between 2000 and 2012, based on classifying nearly 7 million unique patches of observed tree cover gain using optical and radar satellite imagery. Most observed gain patches (69.2%) consisted of small patches of natural regrowth (5.9 ± 0.2 Mha). However, expansion of tree plantations dominated observed increases in tree cover across the tropics (11.8 ± 0.2 Mha) with 92% of plantation expansion occurring in biodiversity hotspots and 14% in arid biomes. We estimate that tree plantations expanded into 9.2% of accessible protected areas across the humid tropics, most frequently in southeast Asia, west Africa, and Brazil. Given international tree planting commitments, it is critical to understand how future tree plantation expansion will affect remaining natural ecosystems. One Sentence Summary: Tree plantations dominated recent expansions of tropical tree cover, including into 9% of accessible parks in the humid tropics.
... Some use only Synthetic Aperture Radar (SAR) (Oon et al., 2019;Cheng et al., 2018b), which is unaffected by the frequent cloud coverage in tropical regions (Cheng et al., 2018a). Perhaps even more popular is the combination of optical and Radar images (Descals et al., 2020;Xu et al., 2020;Laurin et al., 2013;Pohl, 2014;Cheng et al., 2016;Nomura et al., 2019;Sarzynski et al., 2020;Gutiérrez-Vélez and DeFries, 2013). Only a few works map oil palms at country-wide or continental scales. ...
Article
Full-text available
Accurate mapping of oil palm is important for understanding its past and future impact on the environment. We propose to map and count oil palms by estimating tree densities per pixel for large-scale analysis. This allows for fine-grained analysis, for example regarding different planting patterns. To that end, we propose a new, active deep learning method to estimate oil palm density at large scale from Sentinel-2 satellite images, and apply it to generate complete maps for Malaysia and Indonesia. What makes the regression of oil palm density challenging is the need for representative reference data that covers all relevant geographical conditions across a large territory. Specifically for density estimation, generating reference data involves counting individual trees. To keep the associated labelling effort low we propose an active learning (AL) approach that automatically chooses the most relevant samples to be labelled. Our method relies on estimates of the epistemic model uncertainty and of the diversity among samples, making it possible to retrieve an entire batch of relevant samples in a single iteration. Moreover, our algorithm has linear computational complexity and is easily parallelisable to cover large areas. We use our method to compute the first oil palm density map with 10 m Ground Sampling Distance (GSD), for all of Indonesia and Malaysia and for two different years, 2017 and 2019. The maps have a mean absolute error of ±7.3 trees/ha, estimated from an independent validation set. We also analyse density variations between different states within a country and compare them to official estimates. According to our estimates there are, in total, >1.2 billion oil palms in Indonesia covering >15 million ha, and > 0.5 billion oil palms in Malaysia covering >6 million ha.
... Synthetic Aperture Radar (SAR) images have been used widely in forest mapping at both global and regional scales since they can provide cloud-free structural information sensitive to forest cover [5]. Another approach that recent studies have frequently used is the combination of optical and SAR imagery [6]. The common issue that runs through all this research as outlined above is that the focus is on using a subset of variables only. ...
Article
Full-text available
Forests in Sub-Saharan Africa are experiencing some of the highest rates of deforestation and degradation in the world, with most natural forest species being replaced by cropland and plantation monoculture. In this work, a method was developed that combined the Synthetic Aperture Radar (Sentinel-1) and optical satellite imagery (Sentinel-2) data to accurately map natural forest and perennial crops (oil palm) in Ghana. This was done using all three variables including spatial, spectral, and temporal variables to assess the most important variables in characterizing oil palm and natural forest, as well as the added value of sentinel-1 SAR data in a sentinel-2 optical-based classification. In this workflow, the Gray level co-occurrence matrix (GLCM) was calculated as representing textural/spatial variables, a yearly median composite to represent the spectral variables, and raining and dry season composites of Normalized Difference Vegetation Index (NDVI) and Normalized Difference Moisture Index (NDMI) to represent the temporal variables for the Sentinel-2 data. In terms of the SAR data, rainy and dry season composites of NDVI and NDMI were calculated. With all these variables together, a characterization of the study area was conducted based on reference data of the land use land cover classes including oil palm, natural forests, and croplands (others) using Random Forest classifier. The variable importance of the Random Forest model was investigated to identify the top 10 most important variables. Results from this study showed that spectral variables followed by spatial variables are the most important and need to be considered when characterizing oil palm and natural forest, which is consistent with some pieces of literature. The use of sentinel-2 data achieved an acceptable classification accuracy (75%); whereas, sentinel-1 SAR further increased the accuracy (up to 85%) as compared to sentinel-2 only.
... Synthetic Aperture Radar (SAR) images have been used widely in forest mapping at both global and regional scales since they can provide cloud-free structural information sensitive to forest cover [5]. Another approach that recent studies have frequently used is the combination of optical and SAR imagery [6]. The common issue that runs through all this research as outlined above is that the focus is on using a subset of variables only. ...
Article
Full-text available
Forests in Sub-Saharan Africa are experiencing some of the highest rates of deforestation and degradation in the world, with most natural forest species being replaced by cropland and plantation monoculture. In this work, a method was developed that combined the Synthetic Aperture Radar (Sentinel-1) and optical satellite imagery (Sentinel-2) data to accurately map natural forest and perennial crops (oil palm) in Ghana. This was done using all three variables including spatial, spectral, and temporal variables to assess the most important variables in characterizing oil palm and natural forest, as well as the added value of sentinel-1 SAR data in a sentinel-2 optical-based classification. In this workflow, the Gray level co-occurrence matrix (GLCM) was calculated as representing textural/spatial variables, a yearly median composite to represent the spectral variables, and raining and dry season composites of Normalized Difference Vegetation Index NDVI) and Normalized Difference Moisture Index (NDMI) to represent the temporal variables for the Sentinel-2 data. In terms of the SAR data, rainy and dry season composites of NDVI and NDMI were calculated. With all these variables together, a characterization of the study area was conducted based on reference data of the land use land cover classes including oil palm, natural forests, and croplands (others) using Random Forest classifier. The variable importance of the Random Forest model was investigated to identify the top 10 most important variables. Results from this study showed that spectral variables followed by spatial variables are the most important and need to be considered when characterizing oil palm and natural forest, which is consistent with some pieces of literature. The use of sentinel-2 data achieved an acceptable classification accuracy (75%); whereas, sentinel-1 SAR further increased the accuracy (up to 85%) as compared to sentinel-2 only.
Article
Full-text available
Machine learning (ML) offers new technologies in the precision agriculture domain with its intelligent algorithms and strong computation. Oil palm is one of the rich crops that is also emerging with modern technologies to meet global sustainability standards. This article presents a comprehensive review of research dedicated to the application of ML in the oil palm agricultural industry over the last decade (2011–2020). A systematic review was structured to answer seven predefined research questions by analysing 61 papers after applying exclusion criteria. The works analysed were categorized into two main groups: (1) regression analysis used to predict fruit yield, harvest time, oil yield, and seasonal impacts and (2) classification techniques to classify trees, fruit, disease levels, canopy, and land. Based on defined research questions, investigation of the reviewed literature included yearly distribution and geographical distribution of articles, highly adopted algorithms, input data, used features, and model performance evaluation criteria. Detailed quantitative–qualitative investigations have revealed that ML is still underutilised for predictive analysis of oil palm. However, smart systems integrated with machine vision and artificial intelligence are evolving to reform oil palm agri-business. This article offers an opportunity to understand the significance of ML in the oil palm agricultural industry and provides a roadmap for future research in this domain.
Article
Full-text available
Populations of many of Nearctic-neotropical migratory birds have declined in the past several decades, recent estimates suggested a dramatic loss of 2.5 billion birds over the past 50 years in North America. Habitat loss and degradation represent a major threat in the tropics. Managed agroecosystems have the potential to mitigate some impacts of land conversion, however, little is known regarding the habitat quality provided by working landscapes in the overwintering range. In this research, we surveyed the migratory bird community in the rapidly expanding oil palm plantations in southern Mexico; and also the declining population of the Wood Thrush (Hylocichla mustelina) inhabiting forest fragments in an agricultural matrix in Costa Rica. We assessed the value of both human-modified habitats by using a combination of demographic, distributional, and individual habitat quality indicators, as well as the relationship of these indicators with environmental characteristics. In the Mexican oil palm plantations, we found that species richness of migratory birds tended to be higher in forest patches than in oil palm, that community assemblages of migratory birds differed between habitats, and that differences in migratory bird abundance were driven by vegetative structure. Specifically, when differences in indicators occurred between oil palm and native forest, most migratory species exhibited indicators of better habitat quality in the native forest. Lastly, we observed, for the first time, territoriality in oil palm plantations and estimated home range sizes for the American Redstart (Setophaga ruticilla), which tended to be smaller than in the native forest. The Wood Thrush population in Costa Rica exhibited an average territory size estimated of 0.71 ha. We were able to determine associations between fragments' characteristics and body conditions, whereby birds in young and more humid fragments exhibited better fitness. Additionally, fragment size alone is probably not the best indicator of habitat quality for Wood Thrushes in Costa Rica. Our results suggest that most species of migratory birds assessed responded positively to forest structure complexity, and that age and sex ratios combined with measures of the physiological conditions, environmental moisture and home range sizes can be used to assess habitat quality for migratory birds overwintering in working landscapes. Importantly, determining a species’ territoriality dynamics, is key when selecting a given indicator of habitat quality for each species due to distributional behavior. Our results also suggest that management strategies that promote forest-like conditions in oil palm plantations can improve the habitat quality in this agroecosystem for declining populations of migratory birds. Additionally, these findings support potential value in variable-sized forest fragments within agricultural areas for the conservation of the Wood Thrushes, and soil humidity could be used as a proximate cue for food availability and ultimately as a habitat quality indicator. Lastly, our results emphasize the importance of determining territoriality dynamics, assessing various habitat indicators, and long-term monitoring, in order to develop effective management measures to improve the conservation value of working landscapes in the Neotropics to mitigate the high rate of habitat loss and degradation, especially considering that habitat availability in the tropics could be limiting migratory bird populations.
Chapter
Farming in developing countries is majorly dependent on the traditional knowledge of farmers, with unscientific agricultural practices commonly implemented, leading to low productivity and degradation of resources. Moreover, mechanization has not been integral to farming, and thus managing a farm is a time-consuming and labor-intensive process. Consequently, precision agriculture (PA) offers great opportunities for improvement. Using geographic information and communication technology (Geo-ICTs) principles, PA offers the opportunity for a farmer to apply the right amount of treatment at the right time and at the right location in the farm. However, in order to collect timely high-resolution data, drone-based sensing and image interpretation is required. These high-resolution images can give detailed information about the soil and crop condition, which can be used for farm management purposes. Leaf area index, normalized difference vegetation index, photochemical reflectance index, crop water stress index, and other such vegetation indices can provide important information on crop health. Temporal changes in these indices can give vital information about changes in health and canopy structure of the crop over time, which can be related to its biophysical and biochemical stress. These stresses may have occurred due to insufficient soil nutrient, inappropriate soil moisture, or pest attack. Through UAV-based PA, stressed areas can be identified in real time, and some corrective measures can also be carried out (e.g., fertilizer and pesticide spraying). Moreover, the advantages and different approaches to integrate the UAV data in the crop models are also described.
Chapter
As oil palm is cultivated in large-scale plantations, prior and post-planting operations are laborious and time expensive, and also manual implementation of some of these operations often results in incorrect measurements and information. Successful oil palm management in prior and post-planting operations requires effective techniques to collect precise information. Unmanned aerial vehicle (UAV) imagery is a low-cost alternative to field-based assessment but requires the development of methods to easily and accurately extract the required information. The individual oil palm tree detection and height assessment are important and labor-intensive tasks for large-scale plantations where trees taller than 15 m are being replanted due to the high cost of harvesting and low yield. Therefore, in this chapter we demonstrate a general work flow for individual oil palm detection and height assessment using UAV imagery. We explain local maximum (LM) and template matching (TM) techniques as two commonly used approaches in individual tree detection. The accuracy of each method was evaluated on 20 randomly selected plots on UAV image with recall, precision, and F-score method. The F-score for LM method is higher than TM methods which, respectively, are 0.83 and 0.60. From 17,252 oil palm trees in the 20 plots, LM algorithm could detect 1395 (almost 80%) and TM algorithm, 967 (almost 55%). Three hundred fifty-seven and 785 trees have been missed, respectively, in LM and TM approaches. In both cases, background vegetation incorrectly has been labeled as oil palm tree, where 141 and 322 objects have been falsely detected as oil palm tree in LM and TM approaches. LM worked better in almost all of the plots; however the performance decreased in highly dense area. Contradictory to TM approach, in LM approach, shadows did not affect the performance as it was reflected in precision value. In densely cultivated plots due to leaves overlapping of neighboring trees, algorithm failed, and also in sparsely cultivated plots, shadows caused some commission errors. Inherited distortion of UAV image also caused some omission errors in TM approach. Individual detected tree with LM algorithm was used in the next part for oil palm height estimation overlaying with canopy height model.
Article
Full-text available
Continuous oil palm distribution maps are essential for effective agricultural planning and management. Due to the significant cloud cover issue in tropical regions, the identification of oil palm from other crops using only optical satellites is difficult. Based on the Google Earth Engine (GEE), this study aims to evaluate the best combination of open-source optical and microwave satellite data in oil palm mapping by utilizing the C-band Sentinel-1, L-band PALSAR-2, Landsat 8, Sentinel-2, and topographic images, with the Muda River Basin (MRB) as the test site. The results show that the land use land cover maps generated from the combined images have accuracies from 95 to 97%; the best combination goes to Sentinel-1 and Sentinel-2 for the overall classification. Meanwhile, the best combination for oil palm classification is C5 (PALSAR-2 + Landsat 8), with the highest producer accuracy (96%) and consumer accuracy (100%) values. The combination of C-band radar images can improve the classification accuracy of oil palm, but compared with the combination of L-band images, the oil palm area was underestimated. The oil palm area had increased from 2015 to 2020, ranging from 10% to 60% across all combinations. This shows that the selection of optimal images is important for oil palm mapping.
Article
Full-text available
Oil palm (Elaeis guineensis) monocrop has increased worldwide. Plantations have had an impact on tropical landscapes decreasing natural vegetation or replacing other crops. The cultivation of oil palm in Mexico increased and this trend will likely continue. However, there are no documents about the regions where this crop has increased and its impact on the local land use dynamics. This information would help guide public policies. This study had the following objectives: 1) to analyze the trend in the change in the surface where oil palm has been cultivated in Mexico over the last 30 years; and 2) to evaluate changes to land use in municipalities with extensive cultivation surfaces. For this, government data was analyzed and palm plantations in four of the municipalities with major palm plantations were identified. Additionally, Geographic Information Systems were used to conduct a preliminary analysis of the area and of the covers that were replaced. In Mexico, the area in which palm is cultivated has increased seventyfold (from 1318 to 90 118 ha) from 1985 to 2016. Most of those plantations are located in Chiapas, Campeche, Tabasco, and Veracruz. Overall, palm plantations replaced other agricultural systems. In some municipalities, this area exceeded or was the same as the area used for corn (Zea mays) and sorghum (Sorghumn spp.) crops. This might have repercussions on food security. Although on a smaller scale, the expansion of palm crops also boosted the loss and transformation of natural vegetation in some of the municipalities that were part of this study.
Article
Oil palm (Elaies guineensis) plantations are among the fastest growing agroecosystems in the Neotropics, but little is known about how Neotropical birds use oil palm habitats. To better understand the potential value of oil palm as an overwintering habitat for migratory birds, we surveyed birds in oil palm and native forest remnants in Tabasco, Mexico, from 19 December 2017 to 27 March 2018. We collected data on bird abundance and vegetative structure and used generalized linear models and multivariate analysis to assess how oil palm development influenced migrant bird diversity, community assemblages, and abundance. We found that species richness of migratory birds tended to be higher in forest patches than in oil palm, that community assemblages of migratory birds differed between native forest and oil palm plantations, and that differences in migratory bird abundance, and subsequent changes in community assemblages were driven by differences between native forest and oil palm plantations in vegetative structure. The bird community of native forest was characterized by migrant species sensitive to forest loss that forage low in the understory and in the leaf litter, whereas the bird community of oil palm plantations was represented by generalist species that occupy a wider range of foraging niches. Our results suggest that most species of migrant birds responded positively to several forest structural features and that integrating more native trees and increasing the amount of understory vegetation in oil palm plantations may increase the value of working landscapes for migratory birds.
Article
Full-text available
Change assessment is a central and active area of inquiry in remote sensing. Broadly adopted probabilistic methods discriminate between change and no change based on image differencing, normalization and aggregation into a single band metric that is assumed to follow a Chi-square distribution. The adoption of the Chi-square distribution requires the application of band transformation to original data that is computer expensive and that operates under an untested assumption of multivariate distribution of pixel values in input bands. Despite the wide adoption of the Chi-square distribution, its appropriateness for discriminating between change and no change remains an open question. Here, we test the performance of the Chi-square distribution for change assessment compared to the use of the more-generic Gamma distribution. For this purpose, we implement an algorithm that iteratively removes observations labelled as change according to a pre-defined probabilistic distribution and a probability change threshold. We implement the algorithm in three study areas in tropical and subtropical regions representing contrasting ecological conditions and land cover types and changes. We also test whether input multispectral data meets the assumption of multivariate normality required for band transformation and for the use of the Chi-square distribution. We found that the Gamma distribution applied to untransformed data consistently performs more robustly to discriminate between change and no change compared to the application of band transformation and subsequent use of the Chi-square distribution. We also found that, in none of the evaluated cases, input multispectral data meet the assumption of multivariate normality required for band transformation. Our results suggest that assumptions about multivariate normality can affect the robustness of probabilistic change assessment in multispectral remote sensing. We encourage the remote sensing community to adopt the Gamma distribution applied to untransformed data as a probabilistic approach to differentiate between change and no change.
Book
This book showcases how new and emerging technologies like Unmanned Aerial Vehicles (UAVs) are trying to provide solutions to unresolved socio-economic and environmental problems. Unmanned vehicles can be classified into five different types according to their operation. These five types are unmanned ground vehicles, unmanned aerial vehicles, unmanned surface vehicles (operating on the surface of the water), unmanned underwater vehicles, and unmanned spacecraft. Unmanned vehicles can be guided remotely or function as autonomous vehicles. The technology has a wide range of uses including agriculture, industry, transport, communication, surveillance and environment applications. UAVs are widely used in precision agriculture; from monitoring the crops to crop damage assessment. This book explains the different methods in which they are used, providing step-by-step image processing and sample data. It also discusses how smart UAVs will provide unique opportunities for manufacturers to utilise new technological trends to overcome the current challenges of UAV applications. The book will be of great interest to researchers engaged in forest carbon measurement, road patrolling, plantation monitoring, crop yield estimation, crop damage assessment, terrain modelling, fertilizer control, and pest control.
Article
Sustainable oil palm production is a key issue in global biodiversity conservation and sustainable development. As one of the world’s major vegetable oil crops, oil palm has expanded exponentially to meet increased demand over the past decades. However, previous monitoring and assessments of oil palm plantations were hampered because of the lack of high-resolution annual maps at the global scale. We produced annual oil palm plantation maps in 4 major producer countries (Indonesia, Malaysia, Thailand and Papua New Guinea) in Asia-Pacific from 2007 to 2018 at 100-m resolution using advanced remote sensing techniques with Phased Array L-band Synthetic Aperture Radar (PALSAR) and Moderate Resolution Imaging Spectroradiometer (MODIS) data. We uncover the global patterns of oil palm expansion and find that global oil palm expansion has a very high degree of potential conflict with local biodiversity. Globally, 99.9% of oil palm plantations overlapped with Conservation Priority Zones (CPZs) and oil palm plantations encroached on 231 protected areas. We suggest to incorporate the related issues into the Post-2020 Global Biodiversity Framework.
Article
Full-text available
Oil palm (Elaeis guineensis Jacq.) is of the most profitable and widespread commercial high tree crops in the tropical world, typically in Southeastern Asia. The present study aims to provide a brief but broad overview of different applications of expert systems (ESs) in oil palm precision agriculture (PA), focusing on the three main generic categories: crop, water, and soil management. This study is meant to review research articles from the past decade: 2011–2020. Based on the search strategy alongside the inclusion criteria, 108 articles were included for synthesis activity. The findings of the study reveal patterns, networks, relationships, and trends in the application of ESs in oil palm PA in the past decade. The broad insight obtained from the synthesis activity was used to identify the possible roads ahead in oil palm PA. The findings of this study could be useful and beneficial to the research community and stakeholders in identifying the progress and trends of ESs in oil palm PA in the past decade, help to gain a holistic view on research gaps, potential markets, relevant advantages, the roads ahead, and contributing to further systematic research (deepen or broaden) in this topic.
Chapter
Farming in developing countries is majorly dependent on the traditional knowledge of farmers, with unscientific agricultural practices commonly implemented, leading to low productivity and degradation of resources. Moreover, mechanization has not been integral to farming, and thus managing a farm is a timeconsuming and labor-intensive process. Consequently, precision agriculture (PA) offers great opportunities for improvement. Using geographic information and communication technology (Geo-ICTs) principles, PA offers the opportunity for a farmer to apply the right amount of treatment at the right time and at the right location in the farm. However, in order to collect timely high-resolution data, dronebased sensing and image interpretation is required. These high-resolution images can give detailed information about the soil and crop condition, which can be used for farm management purposes. Leaf area index, normalized difference vegetation index, photochemical reflectance index, crop water stress index, and other such vegetation indices can provide important information on crop health. Temporal changes in these indices can give vital information about changes in health and canopy structure of the crop over time, which can be related to its biophysical and biochemical stress. These stresses may have occurred due to insufficient soil nutrient, inappropriate soil moisture, or pest attack. Through UAV-based PA, stressed areas can be identified in real time, and some corrective measures can also be carried out (e.g., fertilizer and pesticide spraying). Moreover, the advantages and different approaches to integrate the UAV data in the crop models are also described.
Article
Article
Full-text available
African oil palm (Elaeis guinensis) is the most productive oil seed. Globally, the oil palm industry plans to double the area under cultivation to meet growing demands for both vegetable oils and biodiesel. Accurate assessment and monitoring of African palm extensification and intensification for both development and sustainability is crucial given that these crops are replacing the natural high-biodiversity forests as well as local subsistence agriculture. Using a simultaneous collection of RADARSAT synthetic aperture radar (SAR) and ground based digital video, we describe and model the spatial distribution of African palm and explore its lifecycle placing it in the regional ecological context of the Ecuadorian, Amazon. We evaluate the strengths and limitations of integrating RADARSAT texture information, Landsat ETM+, and digital video data to distinguish African oil palm plantations from other land-use and land-cover (LULC) categories. The grey-level co-occurrence matrix (GLCM) and a separate hybrid classification approach using a concatenation of SAR-optical products were tested. A significant improvement in the classification accuracy of African palm in the context of the Ecuadorian Amazon was obtained through the fusion of optical and RADARSAT texture measures as compared to single sensor classifications. The fusion of single ETM+ bands with texture measures achieved the highest user’s and producer’s accuracy with 83 percent and 90 percent, respectively.
Conference Paper
Full-text available
Land is a key parameter in Global Environmental Change. The land change science community has for decades focused on the accelerating pressure on the Earth’s limited land resources (e.g. Lambin & Geist, 2006) resulting from contemporary trends in, e.g. globalization, economic wealth, climate change and population increase. Major research efforts have been invested in scrutinizing the proximate and underlying driving forces of land use and land cover changes at local to global scales. Tilman et al. (2001) reported that rapid and widespread agricultural expansion will pose a serious threat to natural ecosystems worldwide over the next 50 years. In addition, Turner et al. (2007) summarized the current state of insight by noting that virtually all land has been affected in some way by human action and that much of this change is a direct consequence of land use: 40% of the Earth’s land surface is used for agriculture (including improved pasture and co-adapted grassland), which accounts for almost 85% of the annual fresh water withdrawal globally. The land use changes have, for example, a major impact on the global carbon budget as well as on biological diversity, and changes in land use strategies are increasingly presented as strategic instruments to counteract climatic changes (e.g. in connection with the Reducing Emissions from Deforestation and Forest Degradation (REDD) scheme or as an argument for promotion of biofuel to replace fossil fuel). Human land use decisions play a crucial role in driving land use (GLP, 2005), but the complexity of the coupled human-environmental land system is widely acknowledged. During the last couple of decades, a general trend has been that local factors are no longer the most significant determinants of agricultural land use decisions. The geographic scales of interaction have changed significantly in recent years. This has major implications for the ways in which we conceptualize and explore the dynamics of global land use in order to enhance our basic understanding of people and the environment they inhabit. To understand these emerging complexities, the notion of land teleconnections has been used to describe the situation in which demands in distant places significantly influence local land uses at the place of production (Haberl et al., 2009; Seto et al., 2010). Palm oil production is a prominent example of one of the few global land uses that have accelerated in importance as opposed to the majority of major agricultural crops, which have remained remarkably constant with regard to production acreage. It is also one of the land uses characterized by teleconnections. Widespread global demands impact on a limited number of local places. During the past few decades, the oil palm has become one of the most rapidly expanding equatorial crops in the world; oil palms are now grown in 43 countries and their total cultivated area accounts for nearly one-tenth of the world’s permanent cropland (Koh & Wilcove, 2008). This impressive and rapid land use alteration caused by palm oil cultivation has been fuelled by the growing demand for vegetable oil on the global market, driven by population growth as well as the general improvement in economic wealth and consumption. The use of palm oil as a biofuel feedstock is still limited, but that may change in the future since palm oil has higher energy efficiency than the current major biofuel crops (soybean and sugarcane). Moreover, the liquid biofuel market is one of the fastest growing markets for agricultural products globally (Gibbs et al., 2008). Oil palm expansion can, however, contribute to deforestation, peat degradation, biodiversity loss, and forest fires and have a range of social implications (Sheil et al., 2009). Hence, oil palm agriculture deserves special attention. Over the past few decades plantations have directly and indirectly caused deforestation (Geist & Lambin, 2001), and oil palm plantations have become a major driver of deforestation in the tropics (Butler et al., 2009; Fitzherbert et al., 2008; Koh & Wilcove, 2009; Koh & Wilcove, 2008). In Malaysia and Indonesia more than half of the oil palm expansion since 1990 has taken place at the expense of forests (Koh & Wilcove, 2008). The present report aims at providing an overview of the magnitude and geographical distribution of oil palm cultivation. It also considers recent trends in the palm oil market and the future prospects for palm oil. By way of background, we briefly summarize the agroecological characteristics of oil palms. The main aim of the paper is, however, to present a quantitative overview of the extent of land transformations related to the global oil palm production.
Article
Full-text available
Mapping and monitoring changes in the distribution of cropland provides information that aids sustainable approaches to agriculture and supports early warning of threats to global and regional food security. Data from synthetic aperture radar (SAR) sensors can make an important contribution to these crop monitoring activities. This study tested the capability of PALSAR multi-polarization and polarimetric data for crop classification. L-Band results were compared with those achieved with a C-Band SAR data set (ASAR and RADARSAT-1), an integrated C- and L-Band data set, and a multi-temporal optical data set (Landsat Thematic Mapper). Using all L-Band linear polarizations corn, soybeans, cereals and hay-pasture were classified to an overall accuracy of 70%. A more temporally-rich C-Band data set provided an accuracy of 80%. Larger biomass crops (corn and soybeans) were well classified using the PALSAR data. C-Band data were needed to accurately classify low biomass crops (cereals and hay-pasture). With a multi-frequency data set an overall accuracy of 88.7% was reached, and many individual crops were classified to accuracies better than 90%. These results were competitive with the overall accuracy achieved using three Landsat images (88.0%). L-Band parameters derived from three decomposition approaches (Cloude-Pottier, Freeman-Durden and Krogager) produced superior crop classification accuracies relative to those achieved using the linear polarizations. Using the Krogager decomposition parameters from all three PALSAR acquisitions, an overall accuracy of 77.2% was achieved. Results reported in this study emphasize the value of polarimetric as well as multi-frequency SAR data for crop classification. With such a diverse capability, a SAR-only approach to crop classification becomes increasingly viable. Access to multi-polarization data from both RADARSAT-2 and TerraSAR-X promises to further advance the use of SAR for agricultural applications.
Article
Full-text available
Article
Full-text available
The Amazon Basin appears poised to experience rapid expansion of oil palm agriculture. Nearly half of Amazonia is suitable for oil palm cultivation, and Malaysian corporations are now moving into the region to establish new plantations while the Brazilian government is considering a law that would count oil palm as "forest" towards a landowner's forest reserve requirement. Strong economic incentives for a major Amazonian oil palm industry are likely, given growing global demands for edible oils, oil-based products, and biofuel feedstocks. We have two main concerns. First, oil palm plantations are ecologically depauperate, supporting little forest-dependent wildlife. Second, we disbelieve political and corporate statements suggesting that oil palm plantations will be concentrated on previously deforested lands in Amazonia. In reality, oil palm producers strongly favor clearing primary forest for plantations because they can reap immediate profits from timber production. These profits subsidize the costs of plantation establishment and maintenance for the initial 3-5 years until the oil palm plantations become profitable. Hence, oil palm agriculture could soon emerge as a major new threat to the Amazonian environment.
Article
Full-text available
This paper investigates the accuracy with which the age since field planting of oil palm (Elaeis guineensis Jacq.) can be estimated from Landsat Thematic Mapper (TM) radiance at pixel and stand scales. The study site, a commercial plantation 30 km south-east of Kuala Lumpur in Selangor, Malaysia, consisted of even-aged blocks from 4 to 21 years old. Spectral data were the six reflective TM bands and three spectral indices. Nonlinear negative relationships between spectral variables and age are compared to published trends in leaf area, stem height and per cent canopy cover for oil palm and other tree plantations. Correlation coefficients between log age and log radiance are moderate and highly significant (p<0.01) for bands 2-5 and 7 (-0.214 to-0.776) at the pixel scale, and increase at the stand scale (r 2=0.985 for log band 5, p<0.01). Relationships are strongest for the mid-infrared bands, especially band 5 (r 2=0.585, p <0.01) and the infrared index (IRI), a normalized difference index of bands 4 and 5 (r 2= 0.48, p<0.01). Direct and inverse linear regression models for log age with log band and log age with IRI squared (IRIsq) were constructed at both scales. Equivalent age was estimated from the models using independent test sets for differing scales and degrees of aggregation of the age classes. Single age classes cannot be estimated accurately at the pixel or stand scales; the lowest RMS error was obtained from the direct model using all bands (RMS error=3.9 years at pixel scale, 2.7 at stand scale). A posteriori aggregation into generalized age classes (<5, 6-10, 11-15, 16-21 years) improved the RMS error but the results were still unacceptably high (2.2, 2.3, 2.7, 6.0 years respectively for direct model 3 using all bands). Acceptable RMS errors down to 0.58 years were obtained for models using IRIsq with generalized age classes developed and applied at the stand scale when variations in ground cover and other variables were averaged out. The spatial pattern of error in equivalent age deserves investigation for precision crop management.
Article
Full-text available
While many reports have been published on radar backscatter characteristics of coniferous and deciduous forests, little work appears to have been done on investigating the backscatter properties of palm trees. In this study, Japanese JERS-1 LHH band, European ERS-1 CVV band and Russian Almaz-1B SHH band SAR data have been acquired over parts of Kedah and Penang states in West Malaysia in order to investigate the radar backscatter properties for oil palms and rubber trees for each of these sensors.Results show that the radar backscatter for the deciduous rubber trees, for both JERS-1 and ERS-1, appear to behave in accordance with what has been reported earlier for coniferous and deciduous trees, that is, scattering on trunks, branches and twigs at L-band and scattering in the canopy at C-band. The JERS-1 backscatter shows limited correlation with the rubber growth while no relation is found in the ERS-1 data.Oil palms with their characteristic structures affect the radar signal differently compared to the situation for rubber trees. Scattering in the large crown is the dominating backscatter mechanism in both the JERS-1 and ERS-1 data. Leaf area index is correlated closest to the backscatter intensity at both bands.Results from the investigation of the Almaz S-band data are rather discouraging, contradicting earlier more positive reports on the usefulness of the sensor. In this study, the forest types and their intermediate growing stages were found to be virtually indistinguishable, including the clear felled areas. These results should however not be attributed to S-band or Almaz data in general, but rather to this particular data set. It is obvious that the quality of Almaz data varies significantly.
Article
Full-text available
This study developed biomass models to calculate carbon stock levels of the West African oil palms (Elaeis guineensis) using multi-date wet and dry season IKONOS images. Two benchmark areas of the derived savanna eco-regions of Africa were selected for analysis. Allometric equations related above-ground palm biomass to their stem heights. Empirical regression models based on field plot data were established to determine wet and dry biomass (kg m) of oil palm plantations in IKONOS images. The best models were exponential, involving bands 3, 3 and 1, or 3 and 4, and explaining between 63 and 72% of the variability in the data. Model evaluations with independent datasets showed there is 28-36% uncertainty in dry biomass predictions. At the landscape level, multi-date IKONOS data mapped oil palm plantations with an overall accuracy of 88-92%. However, the ability of IKONOS data to differentiate various age groups of oil palms was limited with a high degree of intermixing of classes. The best results were obtained when delineating agro-palm (palms mixed with agriculture and fallows), palm of 1-3 years, and palm of 4-5 years at an overall accuracy of 74.5% using all four IKONOS bands. The results indicate the need for additional spectral bands in the IKONOS sensor. The total carbon per unit area of oil palms was calculated across age groups for the two benchmark areas of West Africa and were 14.75 and 14.94 tonnes ha (or Mg ha), respectively. The corresponding dry biomass (kg m) were 29.5 and 29.88 tonnes ha (or Mg ha). The age of the oil palms were between 1 and 5 years across benchmark areas. The mean rate of accumulation of carbon was 2.95 t C ha year in benchmark area 1 and 2.99 t C ha year in benchmark area 2.
Article
Full-text available
Abundant vegetation species and associated complex forest stand structures in moist tropical regions often create difficulties in accurately classifying land-use and land-cover (LULC) features. This paper examines the value of spectral mixture analysis (SMA) using Landsat Thematic Mapper (TM) data for improving LULC classification accuracy in a moist tropical area in Rondonia, Brazil. Different routines, such as constrained and unconstrained least-squares solutions, different numbers of endmembers, and minimum noise fraction transformation, were examined while implementing the SMA approach. A maximum likelihood classifier was also used to classify fraction images into seven LULC classes: mature forest, intermediate secondary succession, initial secondary succession, pasture, agricultural land, water, and bare land. The results of this study indicate that reducing correlation between image bands and using four endmembers improve classification accuracy. The overall classifica- tion accuracy was 86.6% for the seven LULC classes using the best SMA processing routine, which represents very good results for such a complex environment. The overall classification accuracy using a maximum likelihood approach was 81.4%. Another finding is that use of constrained or unconstrained solutions for unmixing the atmospherically corrected or raw Landsat TM images does not have significant influence on LULC classification performances when image endmembers are used in a SMA approach.
Article
Full-text available
This paper describes a comparative evaluation of several speckle reduction and texture analysis techniques, with particular emphasis on their applicability to supervised land cover classification from SAR images. Issues related to suppression of speckle in a uniform area, preservation of edges, and texture preservation are pursued in these filters. Quality of texture features is measured by the relevancy, discriminative power and ease of computation of the features. The discriminative power of texture features is measured using the Jeffreys-Matusita distance and classification performance measured on a validation set independent from the classifier's training set. Classifiers investigated are maximum-likelihood, multi-layer perceptron (MLP) and radial basis function (RBF) neural networks. Classification accuracy is measured by KHAT statistic calculated from confusion matrices. Two SAR images of ERS-1 and E-SAR programme showing different land cover categories within the regions of Douala and Ngaoundere (Cameroon), and a bi-polarized Synthetic Aperture Radar (SAR) image from an agricultural station near the city of Altona (Canada) are used for analysis. Speckle suppression techniques based on the wavelet transform performs the best, followed by the modified K-nearest neighbours and the Lee's local statistic filters. Depending on the nature of the land cover types being classified, texture features derived from second- and third-order histogram performed the best, followed by first-order statistics and features derived using the grey-level difference vector method. Among all classifiers considered, the MLP and the RBF neural networks performed the best, achieving up to 94% overall accuracy for the E-SAR image of Douala, for example.
Article
Full-text available
Land use/land cover (LULC) change occurs when humans alter the landscape, and this leads to increasing loss, fragmentation and spatial simplification of habitat. Many fields of study require monitoring of LULC change at a variety of scales. LULC change assessment is dependent upon high-quality input data, most often from remote sensing-derived products such as thematic maps. This research compares pixel- and object-based classifications of Landsat Thematic Mapper (TM) data for mapping and analysis of LULC change in the mixed land use region of eastern Ontario for the period 1995–2005. For single date thematic maps of 10 LULC classes, quantitative and visual analyses showed no significant accuracy difference between the two methods. The object-based method produced thematic maps with more uniform and meaningful LULC objects, but it suffered from absorption of small rare classes into larger objects and the incapability of spatial parameters (e.g. object shape) to contribute to class discrimination. Despite the similar map accuracies produced by the two methods, temporal change maps produced using post-classification comparison (PCC) and analysed using intensive visual analysis of errors of omission and commission revealed that the object-based maps depicted change more accurately than maximum likelihood classification (MLC)-derived change maps.
Article
Full-text available
High-yield agriculture potentially reduces pressure on forests by requiring less land to increase production. Using satellite and field data, we assessed the area deforested by industrial-scale high-yield oil palm expansion in the Peruvian Amazon from 2000 to 2010, finding that 72% of new plantations expanded into forested areas. In a focus area in the Ucayali region, we assessed deforestation for high- and smallholder low-yield oil palm plantations. Low-yield plantations accounted for most expansion overall (80%), but only 30% of their expansion involved forest conversion, contrasting with 75% for high-yield expansion. High-yield expansion minimized the total area required to achieve production but counter-intuitively at higher expense to forests than low-yield plantations. The results show that high-yield agriculture is an important but insufficient strategy to reduce pressure on forests. We suggest that high-yield agriculture can be effective in sparing forests only if coupled with incentives for agricultural expansion into already cleared lands.
Article
Full-text available
a b s t r a c t This study compiles and analyses national-level data on land use change (LUC) and its causes in Indonesia and Malaysia over the past 30 years. The study also explores the role that palm oil has played in past LUC and that projected growth in palm oil production may play in LUC until 2020 and suggests strategies to minimize negative effects. Data collection for the study revealed that the quality and quantity of data on LUC on a national scale over time are low. Despite these uncertainties, the overview of past LUC indicates that large changes in land use have occurred in Indonesia and Malaysia. In Indonesia, LUC can primarily be characterized by forest cover loss on 40 million ha (Mha) of land, a 30% reduction in forest land. Deforestation in Malaysia has been smaller in both absolute and relative terms, with a forest cover loss of nearly 5 Mha (20% reduction in forest land). Other large changes in Malaysia occurred in permanent cropland (excluding oil palm), which has decreased rapidly since the early 1990s, and in land under oil palm cultivation, which experienced a sharp increase. Projections of additional land demand for palm oil production in 2020 range from 1 to 28 Mha in Indonesia. The demand can be met to a large extent by degraded land if no further deforestation is assumed. In Malaysia, expansion projections range from 0.06 to 5 Mha, but only the lowest projection of oil palm expansion is feasible when only degraded land may be used. The role of palm oil production in future LUC depends on the size of the projected expansion as well as agricultural management factors such as implementation of best management practices, earlier replanting with higher yielding plants, and establishment of new plantations on degraded land. The current use of degraded land needs to be investigated in order to reduce possible indirect LUC, land tenure conflicts, or other social impacts. In addition to minimizing direct and indirect LUC by the palm oil sector, measures that reduce deforestation triggered by other causes must also be implemented. A key element for doing so is better planning and governance of land use, which entails more appropriate demarcation of forest land and protection of land that still has forest cover, improved monitoring of land use, and more research to uncover the complexities and dynamics of the causes and drivers of LUC.
Article
Full-text available
Article
Full-text available
Wildfires have major effects on forest dynamics, succession and the carbon cycle in the boreal biome. They are a significant source of carbon emissions, and current observed changes in wildfire regimes due to changes in climate could affect the balance of the boreal carbon pool. A better understanding of postwildfire vegetation dynamics in boreal forests will help predict the future role of boreal forests as a carbon sink or source. Time series of Normalized Difference Vegetation Index (NDVI) and Normalized Difference Shortwave Infrared Index (NDSWIR) derived from Moderate Resolution Imaging Spectroradiometer (MODIS) aboard the Terra satellite were used to investigate whether characteristic temporal patterns exist for stands of different ages in the Siberian boreal forests and whether their postwildfire dynamics are influenced by variables such as prewildfire vegetation cover. Two types of forests, evergreen needle-leaf (ENF) and deciduous needle-leaf (DNF), were studied by analysing a sample of 78 burned forest areas. In order to study a longer time frame, a chronosequence of burned areas of different ages was built by coupling information on location and age provided by a forest burned area database (from 1992 to 2003) to MODIS NDVI and NDSWIR time series acquired from 2001 to 2005. For each of the burned areas, an adjacent unburned control plot representing the same forest type was selected, with the aim of separating the interannual variations caused by climate from changes in NDVI and NDSWIR behaviour due to a wildfire. The results suggest that it takes more than 13 years for the temporal NDVI and NDSWIR signal to recover fully after wildfire. NDSWIR, which is associated to canopy moisture, needs a longer recovery period than NDVI, which is associated to vegetation greenness. The results also suggest that variability observed in postwildfire NDVI and NDSWIR can be explained partially by the dominant forest type: while 13 years after a fire NDVI and NDSWIR are similar for ENF and DNF, the initial impact appears to be greater on the NDVI and NDSWIR of ENF, suggesting a faster recovery by ENF.
Article
Full-text available
Oil-palm agriculture is the greatest immediate threat to biodiversity in Southeast Asia. Despite the efforts of environmentalists, oil palm continues to expand across the tropics. Those concerned about the impacts of oil palm on biodiversity must face some harsh social, economic, and ecological realities: (i) oil palm has been a very profitable crop; (ii) palm oil is used in so many products that simple, direct actions, such as boycotts, are unlikely to succeed; (iii) there is currently insufficient demand for certified sustainable palm oil and inadequate political clout from environmental groups in two of the biggest markets for palm oil—China and India—to slow the rate of forest conversion; and (iv) oil-palm agriculture has improved the lives of poor rural communities in Southeast Asia (although it has also disenfranchised some indigenous communities). To address the threats posed by oil-palm agriculture to biodiversity, environmentalists must change the behavior of the palm oil business through: (i) regulations to curb undesirable activities (e.g., a ban on converting forests to oil palm); (ii) financial incentives to promote desirable behavior (e.g., production of certified, sustainable oil palm); (iii) financial disincentives designed to discourage undesirable behavior (e.g., consumer pressure on major manufacturers and retailers to use palm oil that does not come from plantations created at the expense of forests); and (iv) the promotion of alternative, more biodiversity-friendly uses of forested land that might otherwise be converted to oil palm. There is no single best approach for dealing with the oil-palm crisis in Southeast Asia; a mixture of regulations, incentives, and disincentives targeted at all sectors of the oil-palm industry is necessary to protect the region’s rapidly disappearing forests. KeywordsTropical deforestation-Oil palm-Monoculture-Biodiversity conservation-Reconciliation ecology-Biofuels
Article
Full-text available
Indonesia has been the world’s largest producer and exporter of palm oil since 2008. This paper discussed the livelihood impacts of oil palm development in Indonesia, based on lessons learnt from Bungo district, in the province of Jambi. The various community-company partnerships that structure the sector are reviewed and the difficulties raised by the joint ventures schemes are discussed. The merits and drawbacks of oil palm as a smallholder crop are then analysed, based on household socio-economic surveys conducted in 2007–2010. The main causes of conflicts between oil palm companies and communities are unclear land tenure, and a recurrent lack of leadership in smallholders’ cooperatives. Under fair partnerships between smallholders and companies, oil palm could become a smallholder friendly crop. The land-use profitability analysis demonstrates the high returns that can be generated by oil palm independent smallholdings, making it highly competitive with rubber, and much more profitable than rice production. KeywordsNucleus Estates and Smallholders scheme-Independent smallholders-Rubber agroforest-Livelihoods impact-Sumatra
Article
Full-text available
The Digital Elevation Model that has been derived from the February 2000 Shuttle Radar Topography Mission (SRTM) has been one of the most important publicly available new spatial data sets in recent years. However, the 'finished' grade version of the data (also referred to as Version 2) still contains data voids (some 836,000 km 2)—and other anomalies—that prevent immediate use in many applications. These voids can be filled using a range of interpolation algorithms in conjunction with other sources of elevation data, but there is little guidance on the most appropriate void-filling method. This paper describes: (i) a method to fill voids using a variety of interpolators, (ii) a method to determine the most appropriate void-filling algorithms using a classification of the voids based on their size and a typology of their surrounding terrain; and (iii) the classification of the most appropriate algorithm for each of the 3,339,913 voids in the SRTM data. Based on a sample of 1304 artificial but realistic voids across six terrain types and eight void size classes, we found that the choice of void-filling algorithm is dependent on both the size and terrain type of the void. Contrary to some previous findings, the best methods can be generalised as: kriging or inverse distance weighting interpolation for small and medium size voids in relatively flat low-lying areas; spline interpolation for small and medium-sized voids in high-altitude and dissected terrain; triangular irregular network or inverse distance weighting interpolation for large voids in very flat areas, and an advanced spline method (ANUDEM) for large voids in other terrains.
Article
Full-text available
This study evaluated segment-based classification paired with non-parametric methods (CART ® and kNN) and interannual, multi-temporal data in the classification of an 11-year chronosequence of Landsat TM/ETM imagery in the Brazilian Amazon. The kNN and CART ® classification methods, with the integration of multi-temporal data, performed equally well in the separation of cleared, re-vegetated, and primary forest classes with overall accuracies ranging from 77 percent to 91 percent, with pixel-based CART ® classifications resulting in significantly lower variance than all other methods (3.2 percent versus an average of 13.2 percent). Segmentation did not improve classification success over pixel-based methods with the used datasets. Through appropriate band selection methods, multi-temporal bands were chosen in 38 of 44 total classifications, strongly suggesting the utility of inter-annual, multi-temporal data for the given classes and region. The land-cover maps from this study allow for an accurate annualized analysis of land-cover and landscape change in the region.
Article
Full-text available
Industrial agricultural plantations are a rapidly increasing yet largely unmeasured source of tropical land cover change. Here, we evaluate impacts of oil palm plantation development on land cover, carbon flux, and agrarian community lands in West Kalimantan, Indonesian Borneo. With a spatially explicit land change/carbon bookkeeping model, parameterized using high-resolution satellite time series and informed by socioeconomic surveys, we assess previous and project future plantation expansion under five scenarios. Although fire was the primary proximate cause of 1989-2008 deforestation (93%) and net carbon emissions (69%), by 2007-2008, oil palm directly caused 27% of total and 40% of peatland deforestation. Plantation land sources exhibited distinctive temporal dynamics, comprising 81% forests on mineral soils (1994-2001), shifting to 69% peatlands (2008-2011). Plantation leases reveal vast development potential. In 2008, leases spanned ∼65% of the region, including 62% on peatlands and 59% of community-managed lands, yet <10% of lease area was planted. Projecting business as usual (BAU), by 2020 ∼40% of regional and 35% of community lands are cleared for oil palm, generating 26% of net carbon emissions. Intact forest cover declines to 4%, and the proportion of emissions sourced from peatlands increases 38%. Prohibiting intact and logged forest and peatland conversion to oil palm reduces emissions only 4% below BAU, because of continued uncontrolled fire. Protecting logged forests achieves greater carbon emissions reductions (21%) than protecting intact forests alone (9%) and is critical for mitigating carbon emissions. Extensive allocated leases constrain land management options, requiring trade-offs among oil palm production, carbon emissions mitigation, and maintaining community landholdings.
Article
Full-text available
Information on the distribution of tropical forests is critical to decision-making on a host of globally significant issues ranging from climate stabilization and biodiversity conservation to poverty reduction and human health. The majority of tropical nations need high-resolution, satellite-based maps of their forests as the international community now works to craft an incentive-based mechanism to compensate tropical nations for maintaining their forests intact. The effectiveness of such a mechanism will depend in large part on the capacity of current and near-future Earth observation satellites to provide information that meets the requirements of international monitoring protocols now being discussed. Here we assess the ability of a state-of-the-art satellite radar sensor, the ALOS/PALSAR, to support large-area land cover classification as well as high-resolution baseline mapping of tropical forest cover. Through a comprehensive comparative analysis involving twenty separate PALSAR- and Landsat-based classifications, we confirm the potential of PALSAR as an accurate (>90%) source for spatially explicit estimates of forest cover based on data and analyses from a large and diverse region encompassing the Xingu River headwaters in southeastern Amazonia. Pair-wise spatial comparisons among maps derived from PALSAR, Landsat, and PRODES, the Brazilian Amazon deforestation monitoring program, revealed a high degree of spatial similarity. Given that a long-term data record consisting of current and future spaceborne radar sensors is now expected, our results point to the important role that spaceborne imaging radar can play in complementing optical remote sensing to enable the design of robust forest monitoring systems.
Article
In two cultivar × density trials for oil palms (Elaeis guineensis) planted in Indonesia, single leaf area, number of green leaves per tree, leaf opening rate per year and rachis length of leaves were followed over fourteen years. The data were analysed to determine the time course of canopy leaf area and to predict the moment of canopy closure. Growth functions were fitted to the observed data. Estimates of leaf area index (L) were based on single leaf area, number of green leaves, leaf opening and planting density. The time course of L was modelled on the basis of the fitted functions to the components. The moment of canopy closure was calculated from the planting density and the functions fitted for rachis length. The modelled time course of L was considerably different from the function fitted to the single leaf area data. The expansion of L was not as rapid as expected from the area growth of single leaves and, after maximum L was reached, a steady decline was observed. The continuously declining number of green leaves was the main cause of these two observations. The time course of L differed considerably between the two experiments. Not only were there large differences in the number of green leaves maintained per tree in the experiments but also the final area of single leaves differed between both experiments. The first factor was a result of the management of the experiments, whereas the second factor was most likely influenced by a difference in soil-related factors at the two locations. Leaf areas and numbers of leaves per tree were different for each cultivar, as was L. This was also the parameter most sensitive to planting density. Individual leaf area and leaf number per tree were not affected by planting density, but rachis length was affected by the planting density treatment. The moment of canopy closure was similar in both experiments. Planting density was the main factor that determined the onset of canopy closure.
Article
The potential of a random forest (RF) classifier for radar-only crop classifications was evaluated for an eastern and western Canadian site. Overall classification accuracies were improved by approximately 4%-5% over traditional boosted decision trees with gains of up to 7% in the accuracies of specific classes. Accuracies above 85% were obtained for key crops including canola, soybeans, corn, and wheat. Variable importance measures generated by the RF classifier showed that the most important acquisitions occurred in late August to early September at peak biomass and after wheat harvest. The least important images were acquired in May and mid-July. The HV and VV polarizations had the most significant contributions, while the HH polarization contributed little throughout the season, except in late September when the HH response was largely driven by soil conditions. The sensitivity of three RF parameters (number of training pixels, number of trees, and number of variables to select from at each split) was evaluated and shown to have negligible influence on overall accuracy. The RF classifier provided large performance gains in terms of processing time relative to the decision tree classifier. The operational potential and implementation considerations for radar-only Canada-wide crop type mapping are discussed in the context of these results.