No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Time-series analysis of multi-resolution optical imagery for quantifying forest cover loss in Sumatra and Kalimantan, Indonesia
Abstract
Monitoring loss of humid tropical forests via remotely sensed imagery is critical for a number of environmental monitoring objectives, including carbon accounting, biodiversity, and climate modeling science applications. Landsat imagery, provided free of charge by the U.S. Geological Survey Center for Earth Resources Observation and Science (USGS/EROS), enables consistent and timely forest cover loss updates from regional to biome scales. The Indonesian islands of Sumatra and Kalimantan are a center of significant forest cover change within the humid tropics with implications for carbon dynamics, biodiversity maintenance and local livelihoods. Sumatra and Kalimantan feature poor observational coverage compared to other centers of humid tropical forest change, such as Mato Grosso, Brazil, due to the lack of ongoing acquisitions from nearby ground stations and the persistence of cloud cover obscuring the land surface. At the same time, forest change in Indonesia is transient and does not always result in deforestation, as cleared forests are rapidly replaced by timber plantations and oil palm estates. Epochal composites, where single best observations are selected over a given time interval and used to quantify change, are one option for monitoring forest change in cloudy regions. However, the frequency of forest cover change in Indonesia confounds the ability of image composite pairs to quantify all change. Transient change occurring between composite periods is often missed and the length of time required for creating a cloud-free composite often obscures change occurring within the composite period itself. In this paper, we analyzed all Landsat 7 imagery with
... One potential way to achieve annual mapping is to use optical earth observation data e.g., Landsat images for the PALSAR gap period (Chen et al., 2018;Shen et al., 2019). However, this requires abundant Landsat images (>4) (Xu et al., 2018a) that are not available in the humid tropical regions and may cause "false changes" and "inter-annual inconsistency" (Broich et al., 2011). Recently, a superresolution mapping method (Li et al., 2017;Qin et al., 2017;Xu et al., 2017) was used to reconstruct missing forest cover change 80 during 2011-2014 (Zhang et al., 2019) by fusing the PALSAR/PALSAR-2 and the MODIS normalized difference vegetation index (NDVI) with dense temporal resolution and phenological information. ...
... Here we aimed to capture an abrupt NDVI changes (breakpoints) in the two given periods, which is assumed to be caused by the conversion of the original land 225 cover type to the oil palm cultivation. Many change detection algorithms and their derivatives have been developed in recent years to detect subtle or abrupt changes in a dense time-series satellite profiles (Broich et al., 2011;Kennedy et al., 2010;Verbesselt et al., 2010b). Most of these algorithms were applied in forest change monitoring and all reach high consistency in detecting significant change (Cohen et al., 2017). ...
... Voluntary zero-deforestation commitments in the palm oil industry were also implemented since 2010 (Focus, 2016). However, how many and to what extent large corporations will pay real 495 attention to the rights of local populations remains unknown (Barr and Sayer, 2012). ...
Increasing global demand of vegetable oils and biofuels results in significant oil palm expansion in Southeast Asia, predominately in Malaysia and Indonesia. The land conversion to oil palm plantations leads to deforestation, loss of biodiversity, and greenhouse gas emission over the past decades. Quantifying the consequences of oil palm expansion requires fine scale and frequently updated datasets of land cover dynamics. Previous studies focused on total changes for a multi-year interval without identifying the exact time of conversion, causing uncertainty in the timing of carbon emission estimates from land cover change. Using Advanced Land Observing Satellite (ALOS) Phased Array Type L-band Synthetic Aperture Radar (PALSAR), ALOS-2 PALSAR-2 and Moderate Resolution Imaging Spectroradiometer (MODIS) datasets, we produced an Annual Oil Palm Area Dataset (AOPD) at 100-meter resolution in Malaysia and Indonesia from 2001 to 2016. We first mapped the oil palm extent using PALSAR/PALSAR-2 data for 2007–2010 and 2015–2016 and then applied a disturbance and recovery algorithm (BFAST) to detect land cover change time-points using MODIS data during the years without PALSAR data (2011–2014 and 2001–2006). The new oil palm land cover maps are assessed to have an accuracy of 86.61 % in the mapping step (2007–2010 and 2015–2016). During the intervening years when MODIS data are used, 75.74 % of the change detected time matched the timing of actual conversion using Google Earth and Landsat images. The AOPD dataset revealed spatiotemporal oil palm dynamics every year and shows that plantations expanded from 2.59 to 6.39 M ha and from 3.00 to 12.66 M ha in Malaysia and Indonesia, respectively (i.e., a net increase of 146.60 % and 322.46 %) between 2001 and 2016. The increasing trends from our dataset are consistent with those from the national inventories, but slightly greater because of inclusion of smallholder oil palm plantations in our dataset. We highlight the capability of combining multiple resolution radar and optical satellite datasets in annual plantation mapping at large extent using image classification and statistical boundary-based change detection to achieve long time-series. The consistent characterization of oil palm dynamics can be further used in downstream applications. The annual oil palm plantation maps from 2001 to 2016 at 100 m resolution is published in the Tagged Image File Format with georeferencing information (GeoTIFF) at https://doi.org/10.5281/zenodo.3467071.
... Although numerous studies have sought to estimate the ages of tropical secondary forests using satellite imagery (e. g. Moran et al., 1994;Kennedy et al., 2010;Achard et al., 2014;Ma et al., 2022), tropical cloud cover and/or low resolution complicate the identification of past images, which are available at irregular intervals, particularly for small areas of forest disturbance (Hansen et al., 2008(Hansen et al., , 2009Broich et al., 2011;Sola et al., 2016). Combining the high-spatialresolution WorldView02 satellite data with Landsat time series data, Fujiki et al. (2016) demonstrated a highly accurate method for dating the early stages of succession after cultivation shifts; however, the results were less accurate in later stages due to image saturation by canopy closure (Castillo et al., 2012;Cao et al., 2015). ...
... In this study, our results clearly revealed that the Δ 14 C method could be used to date tropical secondary forests after disturbance with high accuracy, even in aseasonal tropical rainforest areas. This method is very useful when satellite imagery of a particular secondary forest cannot be found due to tropical-specific cloud cover, or when the area of a secondary forest is too small to be accurately determined from past satellite imagery (Hansen et al., 2008(Hansen et al., , 2009Broich et al., 2011;Sola et al., 2016). It has been shown that when a certain area is disturbed by forest fire or burning and secondary forest succession begins, the age of the largest tree individual in the forest is a very good alternative for estimating the forest age. ...
Following the widespread loss and degradation of old-growth forests throughout the tropics, the area of secondary forests has expanded significantly in the past few decades. Understanding the history of vegetation modification is necessary for the proper maintenance and management of tropical secondary forests. Although satellite image analysis is an effective method for evaluating such changes, observations in tropical regions are limited by dense cloud coverage, and it is difficult to estimate the timings of forest disturbances based on annual rings because the climate is aseasonal and trees do not form clear annual rings near the equator in Southeast Asia. In this study, we developed a new, highly accurate technique for dating tropical secondary forests after disturbances such as shifting cultivation and forest fire in Malaysia based on Δ 14 C content of trees. We established 29 plots (20 m × 20 m) in secondary forests in six areas of Malaysia, mainly in Sarawak, and used time-series satellite images to determine the most recent disturbance with temporal accuracy of <6 years. We collected wood core samples from one or two trees with the largest diameter in each plot and measured the Δ 14 C content of their pith to estimate tree ages. We detected a significant positive correlation between tree age estimated by Δ 14 C dating and the period since the latest disturbance according to the time-series satellite images. The Δ 14 C-based age estimates were approximately five years younger than those obtained by satellite image dating. This difference is considered to indicate the time required for land use shifts after clear cutting or forest fire, and for tree species invasion after land abandonment. Together, our results revealed that the Δ 14 C method may be used to date tropical secondary forests after disturbance with high accuracy, even in aseasonal tropical rainforest areas.
... However, it is difficult to provide continuous cloud-free high-resolution image data in Indonesia due to its geographical conditions that are always covered by clouds. The existence of clouds causes existing objects underneath the cloud to be invisible, so the current information cannot be utilized [4]. Therefore, a high-resolution satellite image mosaic with minimum cloud cover, or even cloud-free, is expected to aid the natural resources and disasters monitoring activities. ...
... The MPB models have similar strengths and weaknesses among each other. In this model, input data can use all archive data without scene selection to minimize cloud cover [4][5][6][7]. Then, it will be processed by the normalization and correction methods such as radiometric correction, TOA, BRDF, and cloud masking correction. ...
High-resolution satellite imageries have been widely used for many applications, including monitoring natural resources and disasters. Unfortunately, the existence of the cloud interferes with these activities. The small coverage of high-resolution satellite imageries is also needed to be extended to a larger coverage for monitoring large areas. A method for generating the cloud-free tile-based mosaic for high-resolution satellite images is proposed in this paper to address the issues, namely Mosaic Tile-Based (MTB). The basic idea of this method is to take advantage of the cloud cover information from multi-temporal high-resolution satellite images to select the best image with the minimum cloud cover. SPOT-6 and SPOT-7 pan-sharpened images are used to test the method’s reliability. First, each image is re-shaped becomes tiles and keeps the information of its acquisition date. Each tile image is then calculated and assessed based on its cloud cover. Tile image selection is conducted in the next step to select the best tile with the lowest percentage of the cloud cover. In the last step, a mosaic of each best tile is conducted to produce a cloud-free mosaic. As a result, the mosaic product has a minimum cloud cover and larger coverage. The proposed method has beneficial as it keeps the original acquisition date on each tile and is faster for image processing activities. Thus, this method can be used as an alternative method for generating cloud-free mosaic for high-resolution satellite images to support natural resources and disaster monitoring.
... Several factors have come together to now allow for systematically mapping land cover at medium to high spatial resolution (10s of metres) across regional (Broich et al., 2011;Ghorbanian et al., 2020;Tulbure & Broich, 2013;Tulbure et al., 2016), national (Griffiths et al., 2019;Jones, 2019;Yang et al., 2020) and global extents (Buchhorn et al., 2020;Chen et al., 2017;GLanCE, 2021;Hansen et al., 2013;Pesaresi et al., 2016;Pickens et al., 2020;Potapov et al., 2021;Worldcover, 2021;Zhang et al., 2020Zhang et al., , 2021. Among the most important of these factors is the availability of freely accessible multi-temporal remote sensing datasets in a user-friendly format as 'Analysis Ready Data', ARD (Claverie et al., 2018;Dwyer et al., 2018;Frantz, 2019;Truckenbrodt et al., 2019;Wulder et al., 2016), such as the Committee on Earth Observation Satellites ARD for Land (CARD4L, https://ceos.org/ard/). ...
... For the reasons outlined above, studies have either used image interpretation of classes of interest, based on spectral-temporal information (Broich et al., 2011;Tulbure et al., 2016) or made use of existing layers . Finally, the assignment of class labels requires regional knowledge to properly label the classes of interest, highlighting the need for regional efforts for such data collection. ...
Unprecedented amounts of analysis‐ready Earth Observation (EO) data, combined with increasing computational power and new algorithms, offer novel opportunities for analysing ecosystem dynamics across large geographic extents, and to support conservation planning and action. Much research effort has gone into developing global EO‐based land‐cover and land‐use datasets, including tree cover, crop types, and surface water dynamics. Yet there are inherent trade‐offs between regional and global EO products pertaining to class legends, availability of training/validation data, and accuracy. Acknowledging and understanding these trade‐offs is paramount for both developing EO products and for answering science questions relevant for ecology or conservation studies based on these data. Here we provide context on the development of global EO‐based land‐cover and land‐use datasets, and outline advantages and disadvantages of both regional and global datasets. We argue that both types of EO‐derived land‐cover datasets can be preferable, with regional data providing the context‐specificity that is often required for policy making and implementation (e.g., land‐use and management, conservation planning, payment schemes for ecosystem services), making use of regional knowledge, particularly important when moving from land cover to actors. Ensuring that global and regional land‐cover and land‐use products derived based on EO data are compatible and nested, both in terms of class legends and accuracy assessment, should be a key consideration when developing such data. Open access high‐quality training and validation data derived as part of such efforts are of utmost importance. Likewise, global efforts to generate sets of essential variables for climate change, biodiversity, or eventually land use, which often require land‐cover maps as inputs, should consider regionalized, hierarchical approaches to not sacrifice regional context. Global change impacts manifest in regions, and so must the policy and planning responses to these challenges. EO data should embrace that regions matter, perhaps more than ever, in an age of global data availability and processing.
... Landsat images for the PALSAR gap period (Chen et al., 2018;Shen et al., 2019). However, this requires abundant Landsat images (> 4; Xu et al., 2018a) that are not available in the humid tropical regions and may cause "false changes" and "inter-annual inconsistency" (Broich et al., 2011). Recently, a super-resolution mapping method (X. ...
... Here we aimed to capture abrupt NDVI changes (break points) in the two given periods, which are assumed to be caused by the conversion of the original land cover type to the oil palm cultivation. Many change-detection algorithms and their derivatives have been developed in recent years to detect subtle or abrupt changes in dense time-series satellite profiles (Broich et al., 2011;Kennedy et al., 2010;Verbesselt et al., 2010b). Most of these algorithms were applied in forest change monitoring, and all reach high consistency in detecting significant change (Cohen et al., 2017). ...
Increasing global demand of vegetable oils and biofuels
results in significant oil palm expansion in southeastern Asia, predominately
in Malaysia and Indonesia. The land conversion to oil palm plantations has posed
risks to deforestation (50 % of the oil palm was taken from forest during
1990–2005; Koh and Wilcove, 2008), loss of biodiversity and greenhouse
gas emission over the past decades. Quantifying the consequences of oil palm
expansion requires fine-scale and frequently updated datasets of land cover
dynamics. Previous studies focused on total changes for a multi-year
interval without identifying the exact time of conversion, causing
uncertainty in the timing of carbon emission estimates from land cover
change. Using Advanced Land Observing Satellite (ALOS) Phased Array type
L-band Synthetic Aperture Radar (PALSAR), ALOS-2 PALSAR-2 and Moderate
Resolution Imaging Spectroradiometer (MODIS) datasets, we produced an annual
oil palm area dataset (AOPD) at 100 m resolution in Malaysia and
Indonesia from 2001 to 2016. We first mapped the oil palm extent using
PALSAR and PALSAR-2
data for 2007–2010 and 2015–2016 and then applied a
disturbance and recovery algorithm (Breaks For Additive Season and Trend – BFAST) to detect land cover change
time points using MODIS data during the years without PALSAR data (2011–2014
and 2001–2006). The new oil palm land cover maps are assessed to have an
accuracy of 86.61 % in the mapping step (2007–2010 and 2015–2016). During
the intervening years when MODIS data are used, 75.74 % of the detected change
time matched the timing of actual conversion using Google Earth and
Landsat images. The AOPD revealed spatiotemporal oil palm dynamics
every year and shows that plantations expanded from 2.59 to 6.39×106 ha and
from 3.00 to 12.66×106 ha in Malaysia and Indonesia, respectively (i.e. a net
increase of 146.60 % and 322.46 %) between 2001 and 2016. The higher
trends from our dataset are consistent with those from the national
inventories, with limited annual average difference in Malaysia (0.2×106 ha)
and Indonesia (−0.17×106 ha). We highlight the capability of combining
multiple-resolution radar and optical satellite datasets in annual
plantation mapping to a large extent by using image classification and
statistical boundary-based change detection to achieve long time series. The
consistent characterization of oil palm dynamics can be further used in
downstream applications. The annual oil palm plantation maps from 2001 to
2016 at 100 m resolution are published in the Tagged Image File Format with
georeferencing information (GeoTIFF) at
https://doi.org/10.5281/zenodo.3467071 (Xu et al., 2019).
... As a result, more methods are developed for retrospective change detection than for change monitoring. Many of these historical change point detection methods were adopt to detect change from satellite image time series (Jamali et al., 2015;Verbesselt et al., 2010b;Broich et al., 2011;Kennedy et al., 2010). In particular, A BFAST (Breaks For Additive Season and Trend, Verbesselt et al., 2010b) method that integrates the EFP and a seasonal-trend model to detect changes in the seasonal and trend components of a time series has been widely applied in vegetation change detection. ...
... Remotely sensed image time series analysis (Verbesselt et al., 2010a;Broich et al., 2011) has been drawing more attention in pixel-based change detection in recent years (Jianya et al., 2008;Banskota et al., 2014) due to the increased availability of long-term satellite image time series and improved computational power. Statistically, these methods can be classified as detecting change in mean , (e.g. by tests based on OLS (Ordinary Least Squares) residuals such as CUSUM (Cumulative Sum) test (Brown et al., 1975)), or change in regression parameters, (e.g. by tests that assess all regression coefficients such as supLM (supremum Lagrange Multiplier) test (Andrews, 1993;Zeileis and Hothorn, 2013)). ...
The massive amount of Earth observations provides important information for
modelling environmental change. The opportunities brought by these multidimensional
data such as global near real-time change modelling bring the
challenge of analysing these data.
The multidimensional Earth observations are usually discretised in a computer
system as arrays, which are a data structure for storing collections of
data that are ordered and indexed in arbitrary dimensions. The goal of this
thesis is to discuss the role of arrays in geoscientific data analysis: How can
they represent spatiotemporal phenomena? What are the array operations and
software implementations and how could they be used to handle geoscientific
data? And, how can array regridding, dimension reduction, and multidimensional
algorithms be applied to analyse geoscientific data? To apply arrays to
solve real-life problems, this thesis attempts to address problems from statistical
spatiotemporal change modelling.
This thesis firstly presents an overview of geoscientific data arrays and
then focuses on developing multidimensional information integration methods
to analyse spatiotemporal change from arrays. The developed spatiotemporal
change modelling methods are motivated by a concrete problem in change
detection using satellite images: analysing a time series of vegetation indices
pixel-wise does not take full advantage of the spatial dependency and all the
reflective spectral information. Two methods were developed and evaluated in
deforestation detection. The first method integrates spatial regression models
to detect time series structural change from spatially independent residuals.
The second method uses principal component analysis to extract useful information
from all the spectral bands for change monitoring.
Our study cases show that and how arrays can contribute to scaling big
geoscientific data analysis and facilitating communication and reproduction of
the data modelling process.
... However, these studies were inhibited by insufficient continuous SAR data and were confined over a short period of time. Freely accessed Landsat imagery enables continuous forest cover updates with moderate spatial resolution from regional to biome scales [20]. In recent years, several global land cover products have been generated from Landsat images (30 m resolution), such as fine resolution observation and monitoring global land cover (FROM-GLC) [21], global forest change (GFC) [3], and GlobeLand30 [22], but timely and long-term datasets of forest changes are still unavailable. ...
... Furthermore, all global products with considerable errors may underestimate or overestimate the forest change at a regional scale [23]. To date, most studies still focus on the mono-temporal precise mapping of forests [20,24], the dynamics of forest changes over a period of decades have rarely been investigated at a regional scale. ...
Dramatic changes of forests have strong influence on regional and global carbon cycles, biodiversity, and ecosystem services. Understanding dynamics of forests from local to global scale is crucial for policymaking and sustainable development. In this study, we developed an updating and object-based image analysis method to map forests in Northeast China using Landsat images from 1990 to 2015. The spatio-temporal patterns of forests were quantified based on resultant maps and geospatial analysis. Results showed that the percentage of forested area occupying the entire northeast China was more than 40%, about 94% of initial forest cover remained unchanged (49.37 × 10 4 km 2) over the course of 25 years. A small net forest loss (1051 km 2) was observed during 1990-2015. High forest gain (10,315 km 2) and forest loss (9923 km 2) both occurred from 2010 to 2015. At the provincial level, Heilongjiang demonstrated the highest rate of deforestation, with a net loss of 1802 km 2 (0.89%). Forest changes along elevation, slope, and distance from settlements and roads were also investigated. Over 90% of forest changes occurred in plains and low mountain areas within the elevation of 200-1000 m and slope under 15 •. The most dramatic forest changes can be found within the distance of 2000 m from settlements and roads. The reclamation of sloping land, construction of settlements and roads, and possible smallholder clearing contributed more to forest loss, while ecological projects and related government policies play an important role on afforestation and reforestation. These results can provide useful spatial information for further research on the driving forces and consequences of forest changes, which have critical implications for scientific conservation and management of forests.
... Dense time series satellite data can provide more detailed information than bi-temporal or annual satellite data [19][20][21][22]. The use of temporally dense satellite data to detect forest disturbances is particularly suitable in the tropics, because rapid vegetation recovery can obscure the signs of disturbance events [23,24]. The enhanced temporal resolution of dense time series satellite data is also important for near-real-time monitoring, as early warnings can provide crucial information for regional forest management, such as the location and timing of illegal logging [25]. ...
... Remote Sens. 2019, 11, Example of disturbance detection in the study area. (a) RGB Landsat 8 data acquired on23 December 2017, (b) disturbance detection using Landsat 8, (c) disturbance detection using Sentinel-1 and (d) disturbance detection using Landsat 8 and Sentinel-1. Disturbance pixels that had greater than 0.5 ha minimum mapping unit were used for visualization. ...
The accurate and timely detection of forest disturbances can provide valuable information for effective forest management. Combining dense time series observations from optical and synthetic aperture radar satellites has the potential to improve large-area forest monitoring. For various disturbances, machine learning algorithms might accurately characterize forest changes. However, there is limited knowledge especially on the use of machine learning algorithms to detect forest disturbances through hybrid approaches that combine different data sources. This study investigated the use of dense Landsat 8 and Sentinel-1 time series data for detecting disturbances in tropical seasonal forests based on a machine learning algorithm. The random forest algorithm was used to predict the disturbance probability of each Landsat 8 and Sentinel-1 observation using variables derived from a harmonic regression model, which characterized seasonality and disturbance-related changes. The time series disturbance probabilities of both sensors were then combined to detect forest disturbances in each pixel. The results showed that the combination of Landsat 8 and Sentinel-1 achieved an overall accuracy of 83.6% for disturbance detection, which was higher than the disturbance detection using only Landsat 8 (78.3%) or Sentinel-1 (75.5%). Additionally, more timely disturbance detection was achieved by combining Landsat 8 and Sentinel-1. Small-scale disturbances caused by logging led to large omissions of disturbances; however, other disturbances were detected with relatively high accuracy. Although disturbance detection using only Sentinel-1 data had low accuracy in this study, the combination with Landsat 8 data improved the accuracy of detection, indicating the value of dense Landsat 8 and Sentinel-1 time series data for timely and accurate disturbance detection.
... Image mosaicing is the process of combining two or more side-lap/overlap images to produce a representative and continuous image that will be used in a further analysis process for an information extraction need. The principle of this image mosaicing is to replace the cloud and haze covered areas with different scene/tile/pixels with the cloud or haze free data (CRISP 2001;Mouginis-mark et al, 2001;Furby 2002;Furby et al, 2006;De Vries et al, 2007;Broich et al, 2011;Ghosh & Kaabouch 2016;Guo et al, 2016;Hansen & Loveland 2012;Roswintiarti et al, 2014;Kustiyo et al, 2015;Kustiyo 2016;Margono et al, 2016). ...
... The area also has a relatively complete object of land cover such as forests, swamps, plantations, shrubs, bushes, paddy fields, settlements, and mangroves. The land cover change of the region is quite dynamic and good for representing an analysis of dynamic land cover changes (Broich et al, 2011;Margono et al, 2014;Setiawan et al, 2015). ...
This paper presents an interoperability of annual tile-based mosaic (MTB) images, as well as a verification of the validity of the model for the time series land cover analysis purposes. The primary data used are MTB image of Landsat-8 of the central part of Sumatra, acquired from January 2015 to June 2017. The method used for the interoperability validation is the digital analysis of three-years time series land cover. The classification was performed with four band spectral groups. Training samples are taken from the image of 2016. The results are then reclassified to improve the overall accuracy score based on Jefferies Matusita (JM) distance. The interoperability can be measured by the average of overall accuracy (AOA) score, namely Good (scores > 80%), Fair (70.0% -79.9%), and Bad (< 70%). The results show that the use of the groups Bands 6-5-4-3-2 performs the consistent accuracy level of Good with an AOA score of 86% for six classes object. Whereas the use of the groups Bands 6-5-4-3-2, Bands 6-5-4, and Bands 6-5 shows the consistent accuracy level of Good up to four classes object with an AOA score of 89%, 82%, and 81%, respectively. It means that the annual mosaic image of MTB model is accepted for the image interoperability with an AOA score of > 80% for six and four classes object. Thus the most efficient for interoperability is the use of Bands 6-5 to analyze four class object of land cover.
... The lack of knowledge on long-term vegetation change may have misled our understanding of the global climate change, and not just that of Indonesia. Satellite-based timeseries analyses examining all available observation with dense periodicity are well-known accurate methods to monitor changes in vegetation for several decades 4,[11][12][13] . The normalized difference vegetation index (NDVI) of the Moderate Resolution Imaging Spectroradiometer (MODIS) observes every part of Earth on a near-daily basis and provides 16-day composite data of the vegetation index, allowing for the display of continuous vegetation changes [14][15][16] . ...
Indonesia has one of the world’s largest tropical forests; thus, its deforestation and environmental degradation are a global concern. This study is the first to perform comprehensive big data analyses with coherent vegetation criteria to measure the vegetation change at a high temporal resolution (every 16-day period) for 20 years and the high administrative resolution (regency or city) all over Indonesia. The normalized difference vegetation index (NDVI) of the Moderate Resolution Imaging Spectroradiometer is analyzed through state space modeling. The findings reveal that the NDVI increases in almost all regencies, except in urban areas. A high correlation between the NDVI change and the time is observed in Sumatra, Papua, and Kalimantan. The gain of the NDVI values is obvious in the Central and Eastern Java Island. Human activities, such as the expansion of agriculture and forestry and forest conservation policies, are the key factors for the observed pattern.
... Recent research has attempted to overcome the limitations of medium-resolution remote sensing data by using methods such as spatial-temporal image fusion (Broich et al., 2011;Cai et al., 2020) and deep learning . However, these methods are not effective for large-scale forest change monitoring. ...
... Debido a su resolución temporal (16 días) y espacial (hasta 30 m), la familia de satélites Landsat ha brindado múltiples casos de éxito para analizar e interpretar fenómenos naturales o impactos ambientales a nivel regional durante más de cuatro décadas (Loveland y Dwyer, 2012). Por ejemplo, Duveiller et al. (2008), Broich et al. (2011) y Ernst et al. (2013 estudiaron zonas tropicales con el interés de monitorear -y evaluar cuando fue posible-deforestación, reforestación, degradación y regeneración de la vegetación. DeVries et al. (2015) desarrollaron un sistema de monitoreo de cambios en la vegetación en algunas selvas de Etiopía, mientras que Zhu et al. (2016) utilizaron imágenes Landsat 5, 7 y 8, para determinar tendencias de verdor en China. ...
The Yucatan Peninsula (YP) is home to 32% of tropical forests of Mexico. Consequently, this area has a high cloudiness throughout the year, which represents a particular challenge for any mid- and long-term plant monitoring study based on satellite-image time series. This paper reports the results of a trend classification analysis of a time series (11 Landsat-7 ETM+ and 150 Landsat 8 OLI images) of the Normalized Difference Vegetation Index (NDVI) by soil and vegetation types in the eastern region of Escarcega, Campeche (YP) from 2014 to 2020. We applied the bfast01 algorithm to classify pixels according to linear trends, either global (a line with a positive or negative slope through the study period) or local (two linear segments, each with a positive or negative slope). The analysis reveals that most of the study region has NDVI values with global linear trends (browning: 47%; greening: 15.39%) and, to a lesser degree, local linear trends (delayed browning: 20.66%; browning to greening: 6.04%; delayed greening: 5.26%; greening to browning: 3.88%) We consider that generalized greening (which pools the greening, delayed greening, and browning to greening classes) and generalized browning (which pools the browning, delayed browning, and greening to browning classes) can be interpreted as dynamics with significant signs of recovery and degradation of the NDVI, respectively. These dynamics were identified mainly in the semi-evergreen medium tropical forest (generalized greening: 10.26%; generalized browning: 25.43%), semi-evergreen low thorny tropical forest (7.66 and 21.76) and the secondary tree vegetation of the medium tropical forest (3.26 and 10.93). The largest areas with any kind of linear local trend were identified in 2017 and 2018.
... Many studies have utilized low to medium resolution optical satellite images (e.g., those acquired by MODIS, ASTER, Landsat, SPOT, and Sentinel-2) to map tree plantations such as rubber (Li and Fox, 2011;Li and Fox, 2011;Liu et al. 2012), oil palm (Broich et al., 2011;Carlson et al., 2013), acacia (Win et al., 2009;Larson, 1993), and bamboo (Vina et al., 2008;Xu et al., 2012). There are many limitations, however, on the use of optical images made it difficult to mapping tree plantation. ...
This project aimed to develop a geodatabase of Industrial Tree Plantations (ITPs) in Caraga Region using Remote Sensing (RS) and Geographic Information System (GIS). The geodatabase is expected to aid in the characterization of ITPs in terms of their types, locations, spatial arrangements, and total area. It also aims to provide a form of documentation of the spatial-temporal aspects of ITP growth and development, and management dynamics. An important part of the geodatabase development is mapping the species types, location and extent of ITPs. The project did this by applying machine learning techniques to available RS datasets and complemented by ground surveys. Another objective of the project is to determine areas suitable for establishing new ITPs through conduct of suitability analysis; and to conduct accessibility analysis of log production flow with the use of geodatabase.
Among the major accomplishments of the project are: (i.) the maps and statistics of ITPs in Caraga Region generated through the analysis of satellite and airborne remote sensing images; (ii.) a PostgreSQL+PostGIS geodatabase of ITPs in the region, including an online geodatabase visualization portal accessible at https://geoitp.ccgeo.info; (iii.) the maps and statistics of areas suitable for ITPs; and (iv.) a characterization and analysis of the spatial location, accessibility, and capability of wood processing plants (WPPs) for log production vis-à-vis existing Falcata plantations in the region.
Aside from the ITP geodatabase, the project has generated a significant number of maps and other data products. For these to be accessible and utilized by the public, these products have been uploaded to the Mindanao Integrated Data Sharing Environment (MInDSEt), an online data portal managed by the Caraga Center for Geo-Informatics, of Caraga Center for Geo-Informatics, Caraga State University, Butuan City, Philippines. Interested users can access the project outputs at http://mindset.ccgeo.info:82/organization/industrial-tree-plantation-itp-research-and
innovation-center).
... In our approach the selected map is unbiased in that the map area matches the unbiased sample-based area estimate, satisfying the good practice guidance (GFOI, 2016) that the map neither underestimates nor overestimates the area of the target class. Examples of previous studies in which this map selection approach has been implemented include mapping wetlands in the Congo (Bwangoy et al., 2010), forest loss in Indonesia (Broich et al., 2011), and soybean cover in the United States (Song et al., 2017) and South America (Song et al., 2021). ...
Forest fires contribute to global greenhouse gas emissions and can negatively affect public health, economic activity, and provision of ecosystem services. In boreal forests, fires are a part of the ecosystem dynamics, while in the humid tropics, fires are largely human-induced and lead to forest degradation. Studies have shown changing fire dynamics across the globe due to both climate and land use change. However, global trends in fire-related forest loss remain uncertain due to the lack of a globally consistent methodology applied to high spatial resolution data. Here, we create the first global 30-m resolution satellite-based map of annual forest loss due to fire. When producing this map, we match the mapped area of forest loss due to fire to the reference area obtained using a sample-based unbiased estimator, thus enabling map-based area reporting and trend analysis. We find an increasing global trend in forest loss due to fire from 2001 to 2019, driven by near-uniform increases across the tropics, subtropical, and temperate Australia, and boreal Eurasia. The results quantify the increasing threat of fires to remaining forests globally and may improve modeling of future forest fire loss rates under various climate change and development scenarios.
... However, it still necessary to explore how practical and effective a Landsat time series will be when constructed through combination of data from these three sensors, and whether the time series will introduce noise interference during forest age estimation. Moreover, the long-term time series available in Landsat data cannot be used at annual intervals for subtropical highlands because clouds usually contaminate the images (Hansen et al. 2008(Hansen et al. , 2009Broich et al. 2011;Sola et al. 2016). Knowing how to remove the noise caused by seasonal changes in forests from the time series is also an important issue that needs to be resolved during forest age estimation. ...
Forest age is significantly correlated with the net primary productivity, biomass, carbon flux, and the community structure of forest ecosystems. A Landsat time series was constructed using archived Landsat data and topographical maps to achieve large-scale spatial data on forest age. An algorithm used to identify forest disturbance based on a Mann-Kendall trend test, Mann-Kendall abrupt change test, and a difference rate index (DRI) was proposed. A forest age estimation scheme was established based on the classification of forest disturbance-recovery scenarios to obtain the spatial distribution data of forest age in the study region. The results show that: (1) through de-clouding and spectral fitting, imagery acquired by the Landsat-5 Thematic Mapper and Landsat-8 Operational Land Imager sensors could be used to construct a Landsat time series over the period of 1987–2018 in subtropical areas with complex topography; (2) a DRI was extracted from the time series as a disturbance indicator, which was subjected to a Mann-Kendall trend test, leading to the identification of five forest disturbance-recovery scenarios: recovery (or no recovery) after complete disturbance, recovery after partial disturbance, sustained recovery after positive disturbance, and non-disturbance; (3) based on identification of disturbance-recovery scenarios, a forest age estimation scheme was further developed by using the mean fractional vegetation cover before disturbance, fractional vegetation cover at the end of disturbance, and the vegetation recovery rate after disturbance in conjunction with Landsat Multispectral Scanner data from 1974 and topographical maps from the 1960 s, which achieved overall accuracy metrics of R2=0.72 and RMSE=7.8 years for forest age estimates. Specifically, the accuracy of forest age estimates was high in middle-aged and near-mature forests but low in young and mature forests, regardless of the forest vegetation type. The proposed algorithm for identifying areas of forest disturbance and forest age estimation can allow for forest change monitoring and forest age estimation at a regional scale of subtropical mountainous areas, providing a reference for the remote sensing estimation of forest ecological parameters in those areas.
... Forest conversion impacts on hydrological function mainly through the transformation of rainfall into surface runoff [1]. Most of the changes in forest cover in Indonesia are conversion to plantations and agricultural land [2,3]. Although economically profitable, uncontrolled forest conversion will have a negative impact on ecological functions [4]. ...
It has become a public opinion that forests will always reduce the amount of surface runoff compared to other types of land cover. However, previous research reports that the runoff generalization process is still not fully explained. The activity of converting forests into agricultural land will also have an impact on surface runoff. This study aims to compare the magnitude of surface runoff on the plot scale between forest and coffee combination cassava fields. The surface runoff measurement plot with a size of 8 x 10 meters in the direction of the slope. The higher average surface runoff occurs in a forest plot that is equal to 0.286 ± 0.438 mm, while coffee and cassava plots produce an average surface flow of 0.022 ± 0.057 mm. Understorey and litter are influential in generalizing runoff in both plots. The understorey vegetation cover in the forest plot is lower than in the coffee combination cassava plot. This research confirms the results of previous studies that forests with less understorey vegetation conditions and less litter cover the soil surface can still produce high runoff results.
... Decisions trees are popular supervised classification methods for wetland applications that also have applicability to a wide range of other data problems such as ranking, probability estimation, regression, and clustering and to a variety of remote sensing land cover mapping applications, including wetlands [276][277][278][279][280]. Decision trees are based on a series of logical decisions that are easily interpreted; however, their high expressivity results in a tendency to overfit models [281]. ...
The Prairie Pothole Region (PPR) of North America is an extremely important habitat for
a diverse range of wetland ecosystems that provide a wealth of socio-economic value. This paper describes the ecological characteristics and importance of PPR wetlands and the use of remote sensing for mapping and monitoring applications. While there are comprehensive reviews for wetland remote sensing in recent publications, there is no comprehensive review about the use of remote sensing in the PPR. First, the PPR is described, including the wetland classification systems that have been used, the water regimes that control the surface water and water levels, and the soil and vegetation characteristics of the region. The tools and techniques that have been used in the PPR for analyses of geospatial data for wetland applications are described. Field observations for ground truth data are critical for good validation and accuracy assessment of the many products that are produced. Wetland classification approaches are reviewed, including Decision Trees, Machine Learning, and object versus pixel-based approaches. A comprehensive description of the remote sensing systems and data that have been employed by various studies in the PPR is provided. A wide range of data can be used for various applications, including passive optical data like aerial photographs or satellite-based, Earth-observation data. Both airborne and spaceborne lidar studies are described.
A detailed description of Synthetic Aperture RADAR (SAR) data and research are provided. The state of the art is the use of multi-source data to achieve higher accuracies and hybrid approaches. Digital Surface Models are also being incorporated in geospatial analyses to separate forest and shrub and emergent systems based on vegetation height. Remote sensing provides a cost-effective mechanism for mapping and monitoring PPR wetlands, especially with the logistical difficulties and cost of field-based methods. The wetland characteristics of the PPR dictate the need for high resolution in both time and space, which is increasingly possible with the numerous and increasing remote sensing
systems available and the trend to open-source data and tools. The fusion of multi-source remote sensing data via state-of-the-art machine learning is recommended for wetland applications in the PPR. The use of such data promotes flexibility for sensor addition, subtraction, or substitution as a function of application needs and potential cost restrictions. This is important in the PPR because of the challenges related to the highly dynamic nature of this unique region.
... Salah satu kebijakan pemerintah yang paling dominan dalam memengaruhi perubahan bentang alam di luar jawa adalah program transmigrasi karena para transmigran mengkonversi hutan alam menjadi areal pertanian dan perkebunan (Syam et al. 1997;Miyamoto 2006). Spontaneous transmigrant yang tidak terekam oleh pemerintah juga menjadi penyebab utama pembukaan hutan menjadi areal pertanian dan perkebunan (Feintrenie et al. 2010b;Broich et al. 2011b;Margono et al. 2012 Pembukaan hutan dilakukan sebagian besar dilakukan oleh transmigran, tetapi suku asli juga berkontribusi dalam pembukaan lahan. Hal ini disebabkan karena adanya kecemburuan sosial suku asli terhadap transmigran yang mengelola lahan, sehingga suku asli yang berprofesi sebagai nelayan ikut serta dalam membuka hutan sebagai bentuk klaim atas lahan. ...
As a country considered as land-based economic resources, Indonesia has shown that the expansion of plantation areas and Industrial Plantation Forests (HTI) in this country is growing rapidly. These sectors provide economic solutions and fulfill the market needs. However, environmental degradation and social conflicts caused by the land changes are against the goal of sustainability. The use of the land itself is a manifestation of power competition between actors with an interest in land. In the process of land allocation, various actors with an interest in land will compete with the power they have. The most powerful actors will take control of the land-use based on their decisions. Therefore, to obtain a comprehensive understanding of Indonesia's land cover and land-use changes, identification is needed to reveal the trends. It will also show how big the land cover and land-use changes, the land-use conditions at the site level as well as the understanding of the strong actors determining the process. This study has a general objective to explain land-use and land cover changes spatially and to describe the dynamics of the actors’ power in the land-use for oil palm and HTI plantations. These research objectives can be achieved through intermediate objectives namely: (1) Understanding the process of land cover and land-use change and its actors in an empirical case study; (2)Produce a map of land-use change from 1990-2019; (3) Describing the powerdynamics of actors through ACP and SNA approach and identify the most powerfulactors.
This study was conducted in three stages; systematic review, spatial analysis, analysis of the actors’ strength and social networks. This systematic review is an approach to determine the limits of existing knowledge so that further research builds on that knowledge. This study used a spatial approach to determine the locations that have experienced changes, the extent to which changes in the landscape have occurred, transitions of land-use changes, and the most critical changes in a certain period. It aims to understand the magnitude of the impact caused by the actor and what conditions affect the land change over a certain period. Image interpretation carried out in this study is visual interpretation (digitization on-screen) to classify land cover into several classes and perform accuracy tests based on the coordinates taken in the field. The final stage of this research is the analysis of the actors’ power and their networks by combining the Actor Centered Power (ACP) and Social Network Analysis (SNA) approaches. The combination of these approaches is a form of developing an analysis of the actors’ power in contesting the use of land resources. The case study was conducted in Bengkalis Island, Riau. Actors and their relationships were obtained through literature studies, semi-structured interviews with snowball sampling techniques, observation, and triangulation.
The results of this study indicate that the direct causes of landscape changes in Sumatra and Kalimantan are dominated by oil palm plantation expansion, timber extraction/logging, and HTI expansion. This condition is affected by institutional and policy factors produced by the government. The issue of land ownership and
the weakness of government institutions in carrying out its role as the highest hierarchy in control of land governance are the major causes of the uncontrolled landscape changes in the two regions. All aspects that cause landscape change are the result of the roles of the actors in it. Local and national governments are the actors that most contributed to landscape change through the policies and decisions they make. Farming communities (both indigenous and non-native) and companies are the actors who mostly carry out activities that directly cause changes in the landscape. On Bengkalis Island, a very significant decrease in forest cover began in 1990, followed by an increase in community-managed mixed gardens and oil palm plantations, both managed by large-scale companies and independent companies. On 2019, the forest cover on the island, which is one of the Hydrological Peat Areas (KHG), was only left to 10% of the total island area.
The institutional problems in this area have led to the formation of informal networks for land management. Based on this finding, the most powerful actors in the land-use contestation process came from the actors at the site level, not the central government who holds the highest authority in controlling land governance. The site actors namely farmer activists and village officials, have been the most powerful actors in two different periods. This is based on the value obtained from the analysis of the two actors. Farmer activists have the highest eigenvector value and have the potential to be leaders. Meanwhile, the village officials with the highest betweenness values were found to be manipulating information to develop oil palm plantations in smallholder management areas that overlap with HTI concessions. SNA is a robust framework for developing ACP theoretical frameworks in analyzing actor’s power. SNA can explain several concepts that could not be covered by ACP, including: (1) In comparison to strong ties, the weak ties could encourage the formation of collective action because these ties encourage broader relationships with more diverse actor (the strength of weak ties), (2) SNA describes that an actor can use his power against other actors without interacting directly (action at a distance); (3) SNA can describe that the source of power between one actor and another is interlinked.
To improve the condition of land governance by using a network perspective, the governments as decision-makers and generate policy should understand social network analysis in policy processes. This understanding serves to produce policy interventions that focus on central actors. Farmer activists are central actors who have the potential to become opinion leaders in a collaborative forum to improve land governance. In contrast to the village officials who play a role as intermediaries, while taken advantage of their position to manipulate information through the issuance of land legality documents need to be coerced by actors who develop and implement laws in Indonesia.
Keywords: Actor Centered Power, land governance, spatial analysis, Social Network Analysis, systematic review
... The reason for this is that most of the existing classification technologies only focus on single temporal images and do not make full use of the temporal information of multitemporal images, so the improvement of classification accuracy is limited. Therefore, the multitemporal classification method will be an important research direction of the new classification strategy [16][17][18][19][20]. There are two kinds of land cover classification methods considering temporal context information. ...
Land cover is of great significance for the study of global ecological environmental change. Multitemporal land cover can help us to understand the change process of the regional environment and formulate corresponding protection policies. For single-period image classification, the spatial-temporal information is often ignored, and the classification accuracy is difficult to improve. In this paper, an iterative hidden Markov model (STHMM) is proposed to optimize the multitemporal classification, in which a stacked autoencoding classifier is used to calculate the initial class probability, and the extended random walker-based spatial optimization technique is adopted to optimize the class probability. Finally, the hidden Markov model with expectation maximization is built by exploiting postprocessing temporal optimization. Experimental results show that the proposed method can outperform other classification techniques, and the spatial-temporal hidden Markov model proposed in this paper exhibits more stable and reliable performance and can be widely used in multitemporal classification.
... In recent decades, remote sensing techniques applied in forestry has been given an increased attention, with the ability to monitor the changing pattern of forest cover over a period of time up to extracting important information for forest planning and sustainable management such as forest structure, composition and volume growth (Shao, 2012). Most conservation researchers and practitioners currently rely on satellite-based remote sensing for mapping and monitoring land use change (Broich et al., 2011). Optical remote sensing imagery has been to a paradigm shift in the decade from freely available (e.g. ...
Lately, the issue of forest destruction caused by various factors has received serious
attention from various parties. This is because forest plays an important role in ensuring overall ecosystem stability. Therefore, forest managers should continuously seek forest information quickly, accurately and cost-effectively. Difficulty faced in detecting forest cover changes using conventional method due to inaccessible sites and complexity of Malaysia tropical forest landscape. Conventional field verification of forest cover changes are costly and time consuming. Some of the forest changes are very small and undetectable even though using high resolution satellite imagery. Hence, the usage of UAV equipped with high resolution sensor for data acquisition at the area of interest (AOI) is needed. The focus of this study is to monitor and assess deforestation and forest degradation In Sungai Menyala Forest Reserve, Negeri Sembilan through
mapping and analysis of forest cover changes using geospatial approaches. Early information on forest cover change is extracted from Forest Monitoring Using Remote Sensing System (FMRS) equipped with SPOT 6/7 satellite imageries, forestry and secondary data. Supervised classification of multi-temporal images was carried out. Meanwhile, data acquired from UAV at the AOI is mosaic and analyzed. Next, resultant classes changes is compared and verified for ground truthing and verification. Therefore, the satellite images map at the selected area can be provide to the authorities for more effective and efficient forest management.
... In this context, remote sensing can be put forward as a fundamental approach for mapping and monitoring essential biodiversity variables on both a global and local scale [34][35][36][37]. Regarding vegetation mapping, the most common approach to classifying land-cover types and monitoring vegetation dynamics is currently satellite-based remote sensing [28,[38][39][40][41], followed by piloted aircraft imagery [4,42,43]. However, the pixel size range (e.g., 5 to 30 m) of high-resolution satellite data (e.g., Pleiades, Sentinel, Landsat) hinders the detection of small-scale changes or the differentiation between similar land-cover types. ...
Woody species encroachment on grassland ecosystems is occurring worldwide with both negative and positive consequences for biodiversity conservation and ecosystem services. Remote sensing and image analysis represent useful tools for the monitoring of this process. In this paper, we aimed at evaluating quantitatively the potential of using high-resolution UAV imagery to monitor the encroachment process during its early development and at comparing the performance of manual and semi-automatic classification methods. The RGB images of an abandoned subalpine grassland on the Western Italian Alps were acquired by drone and then classified through manual photo-interpretation, with both pixel-and object-based semi-automatic models, using machine-learning algorithms. The classification techniques were applied at different resolution levels and tested for their accuracy against reference data including measurements of tree dimensions collected in the field. Results showed that the most accurate method was the photo-interpretation (≈99%), followed by the pixel-based approach (≈86%) that was faster than the manual technique and more accurate than the object-based one (≈78%). The dimensional threshold for juvenile tree detection was lower for the photo-interpretation but comparable to the pixel-based one. Therefore, for the encroachment mapping at its early stages, the pixel-based approach proved to be a promising and pragmatic choice.
... We developed predictive models of percent forest canopy cover from 1992 to 2017 using training data derived by sampling an existing forest cover layer obtained from the Google Earth Engine that was developed from Landsat satellite imagery [3], for the time period from 2000 to 2015 ( Figure 2). Existing validation data for forest cover losses for Cambodia were unavailable from [3], but comparisons with expert-delineated field data showed coefficients of determination (R 2 ) of 0.65 and 0.67 for nearby Indonesia [48]. We developed statistical models to predict forest cover for the 1990s as well as 2016 and 2017, which were time periods that were not included in the [3] training and validation dataset. ...
The Mekong River is a globally important river system, known for its unique flood pulse hydrology, ecological productivity, and biodiversity. Flooded forests provide critical terrestrial nutrient inputs and habitat to support aquatic species. However, the Mekong River is under threat from anthropogenic stressors, including deforestation from land cultivation and urbanization, and dam construction that inundates forests and encourages road development. This study investigated spatio-temporal patterns of deforestation in Cambodia and portions of neighboring Laos and Vietnam that form the Srepok–Sesan–Sekong watershed. A random forest model predicted tree cover change over a 25-year period (1993–2017) using the Landsat satellite archive. Then, a statistical predictive deforestation model was developed using annual-resolution predictors such as land-cover change, hydropower development, forest fragmentation, and socio-economic, topo-edaphic and climatic predictors. The results show that almost 19% of primary forest (nearly 24,000 km2) was lost, with more deforestation in floodplain (31%) than upland (18%) areas. Our results corroborate studies showing extremely high rates of deforestation in Cambodia. Given the rapidly accelerating deforestation rates, even in protected areas and community forests, influenced by a growing population and economy and extreme poverty, our study highlights landscape features indicating an increased risk of future deforestation, supporting a spatial framework for future conservation and mitigation efforts.
... For example, due to palm oil plantation expansion, the current forest cover in Kalimantan of Indonesia declined from 75% in the mid-1980s, with an annual deforestation rate of 1.3 million ha. As a result, Kalimantan suffered the highest rate of deforestation in Indonesia [40,41]. The drivers of such large-scale of deforestation are economic, social, and especially from the decentralization policy implemented in Indonesia since 2000. ...
Research Highlights: Our findings highlight that the contribution of carbon sequestration from plantations to REDD+ will remain limited, and that opportunity costs in Southeast Asia will likely increase, due to future oil palm expansion. Background and Objectives: Land use, land-use change, and forestry (LULUCF) are significant sources of carbon emissions. The United Nations Framework Convention on Climate Change (UNFCCC) agreed that the Reducing Emissions from Deforestation and Forest Degradation Plus program, also known as REDD+, could contribute to carbon sinks in tropical regions. These reductions could serve as carbon credits that offset emissions from other sources. Materials and Methods: This study uses the cellular automaton technique to simulate the business-as-usual (BAU) scenario and the gain-loss method, to measure carbon emissions resulting from forest conversion. The output of the integration of the models makes it possible to evaluate one of the most important financial costs: opportunity costs. Two scenarios (with and without consideration of carbon sequestration) in rubber and oil palm plantations are examined. Results: A sensitivity assessment in Kalimantan, Indonesia, shows that carbon sequestration from plantations affects value of opportunity costs less than social discount rates. Further analysis suggests that oil palm plantations have a greater impact than rubber plantations. Conclusions: Our study provides a case that can be applied to other regions for evaluating the impacts of plantation carbon sequestration, and insights that can help local policymakers design a financially attractive REDD+ program in other forest areas of the world.
... Once a decision tree is formulated, external (non-training) data are run through the tree, adhering to its splitting criteria at each node until it reaches the set impurity threshold for that particular decision criteria at the 'leaf' level, thereby yielding a class prediction. A variety of decision tree algorithms such as Classification Tree Analysis, Stochastic Gradient Boosting and Classification and Regression Tree [213][214][215] have been applied to numerous remote sensing land cover applications, including wetlands [216][217][218][219][220]. Baker et al. [213] noted Stochastic Gradient Boosting to be preferable to Classification Tree Analysis for mapping wetland, non-wetland and riparian land cover classes. ...
The following review is the second part of a two part series on the use of remotely sensed data for quantifying wetland extent and inferring or measuring condition for monitoring drivers of change on wetland environments. In the first part, we introduce policy makers and non-users of remotely sensed data with an effective feasibility guide on how data can be used. In the current review, we explore the more technical aspects of remotely sensed data processing and analysis using case studies within the literature. Here we describe: a) current technologies used for wetland assessment and monitoring; b) the latest algorithmic developments for wetland assessment; c) new technologies; and d) a framework for wetland sampling in support of remotely sensed data collection. Results illustrate that high or fine spatial resolution pixels (≤10 m) are critical for identifying wetland boundaries and extent, and wetland class, form and type, but are not required for all wetland sizes. Average accuracies can be up to 11% better (on average) than medium resolution (11-30 m) data pixels when compared with field validation. Wetland size is also a critical factor such that large wetlands may be almost as accurately classified using medium-resolution data (average = 76% accuracy, stdev = 21%). Decision-tree and machine learning algorithms provide the most accurate wetland classification methods currently available, however, these also require sampling of all permutations of variability. Hydroperiod accuracy, which is dependent on instantaneous water extent for single time period datasets does not vary greatly with pixel resolution when compared with field data (average = 87%, 86%) for high and medium resolution pixels, respectively. The results of this review provide users with a guideline for optimal use of remotely sensed data and suggested field methods for boreal and global wetland studies.
... of an index such as the NDVI, a statistical measure such as a mean or median, and so on [54][55][56][57][58]. Creation of composite layers based on monthly/seasonal/annual or statistical measures has shown great potential for characterizing land-cover classes [59][60][61][62][63]. Because these composite layers are extracted from time-series of observations, they could provide higher levels of consistency compared to single-time imagery. ...
Knowledge of land cover and land use nationally is a prerequisite of many studies on drivers of land change, impacts on climate, carbon storage and other ecosystem services, and allows for sufficient planning and management. Despite this, many regions globally do not have accurate and consistent coverage at the national scale. This is certainly true for Ethiopia. Large-area land-cover characterization (LALCC), at a national scale is thus an essential first step in many studies of land-cover change, and yet is itself problematic. Such LALCC based on remote-sensing image classification is associated with a spectrum of technical challenges such as data availability, radiometric inconsistencies within/between images, and big data processing. Radiometric inconsistencies could be exacerbated for areas, such as Ethiopia, with a high frequency of cloud cover, diverse ecosystem and climate patterns, and large variations in elevation and topography. Obtaining explanatory variables that are more robust can improve classification accuracy. To create a base map for the future study of large-scale agricultural land transactions, we produced a recent land-cover map of Ethiopia. Of key importance was the creation of a methodology that was accurate and repeatable and, as such, could be used to create earlier, comparable land-cover classifications in the future for the same region. We examined the effects of band normalization and different time-series image compositing methods on classification accuracy. Both top of atmosphere and surface reflectance products from the Landsat 8 Operational Land Imager (OLI) were tested for single-time classification independently, where the latter resulted in 1.1% greater classification overall accuracy. Substitution of the original spectral bands with normalized difference spectral indices resulted in an additional improvement of 1.0% in overall accuracy. Three approaches for multi-temporal image compositing, using Landsat 8 OLI and Moderate Resolution Imaging Spectroradiometer (MODIS) data, were tested including sequential compositing, i.e., per-pixel summary measures based on predefined periods, probability density function compositing, i.e., per-pixel characterization of distribution of spectral values, and per-pixel sinusoidal models. Multi-temporal composites improved classification overall accuracy up to 4.1%, with respect to single-time classification with an advantage of the Landsat OLI-driven composites over MODIS-driven composites. Additionally, night-time light and elevation data were used to improve the classification. The elevation data and its derivatives improved classification accuracy by 1.7%. The night-time light data improve producer’s accuracy of the Urban/Built class with the cost of decreasing its user’s accuracy. Results from this research can aid map producers with decisions related to operational large-area land-cover mapping, especially with selecting input explanatory variables and multi-temporal image compositing, to allow for the creation of accurate and repeatable national-level land-cover products in a timely fashion.
... We used TOA data products for both Landsat 8 and Sentinel-2 to maintain some degrees of comparability; however, adding an extra step in the framework to calculate surface reflectance data product would result in more accurate estimates of biophysical characteristics of land covers. Previous studies have shown the potential of combining the visible, near infrared (NIR) and shortwave infrared (SWIR) bands from the Landsat suite for identifying forest degradation and forest cover change across a range of ecosystem (Broich et al., 2011;Hansen et al., 2008;Potapov et al., 2009Potapov et al., , 2014Potapov et al., , 2015. Varying the native spatial resolution by the method of pan-sharpening is also known to provide better accuracy in such studies (Meroni et al., 2017). ...
Sustainable Development Goal (SDG) indicator 15.1.1 proposes to quantify "Forest area as a proportion of total land area" in order to achieve SDG target 15.1. While area under forest cover can provide useful information regarding discrete changes in forest cover, it does not provide any insight on subtle changes within the broad vegetation class, e.g. forest degradation. Continental or national-level studies, mostly utilizing coarse-scale satellite data, are likely to fail in capturing these changes due to the fine spatial and long temporal characteristics of forest degradation. Yet, these long-term changes affect forest structure, composition and function, thus ultimately limiting successful implementation of SDG targets. Using a multi-scale, satellite-based monitoring approach, our goal is to provide an easy-to-implement reporting framework for South Asian forest ecosystems. We systematically analyze freely available remote sensing assets on Google Earth Engine for monitoring degradation and evaluate the potential of multiple satellite data with different spatial resolutions for reporting forest degradation. Taking a broad-brush approach in step 1, we calculate vegetation trends in six south Asian countries (Bangladesh, Bhutan, India, Nepal, Pakistan, and Sri Lanka) using the Moderate Resolution Imaging Spectroradiometer (MODIS) Normalized Difference Vegetation Index (NDVI) during 2000-2016. We also calculate rainfall trends in these countries using the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) rainfall data, and further calculate Rain-Use Efficiency (RUE) that shows vegetation trends in the context of rainfall variability. In step 2, we focus on two protected area test cases from India and Sri Lanka for evaluating the potential of finer-resolution satellite data compared to MODIS, i.e. Landsat 8, and Sentinel-2 data, for capturing forest degradation signals, which will ultimately contribute towards SDG indicators 15.1.1 and 15.1.2. We find that most countries show a fluctuating trend in vegetation condition over the years, along with localized greening and browning. The Random Forest (RF) classifier utilized in step 2 was able to generate accurate maps (87% and 91% overall accuracy for Indian and Sri Lankan test cases, respectively) of non-intact forest within the protected areas. We find that almost one-third of the Indian test case is degraded forest, even though it shows overall greening as per the broad-brush approach. This finding corroborates our argument that utilizing higher-resolution satellite data (e.g. 10-m) than those normally used for national-level studies will be crucial for reporting SDG indicator 15.2.1: "progress towards sustainable forest management".
... This makes sense because most of Canada has a relatively rich data stack with only some data-sparse regions (Wulder et al., 2016;Hermosilla et al., 2019); In data-sparse situations, this approach has the same effect as multi-year compositing. As noted previously in Panama (Pelletier et al., 2012) and in other cloudy areas in the tropics (Asner, 2001;Broich et al., 2011;Hansen et al., 2016), the time interval required to construct a composite strongly confounds the ability to detect change. Forest-change mapping efforts in other sparse data situations in the tropics have compensated by including images from all seasons and dampening phenological noise with median scores (Dutrieux et al., 2016;Santos et al., 2016). ...
Measuring and progressing toward international goals of curbing deforestation and improving livelihoods of people who depend on forests requires nuanced understanding of forests and the processes surrounding deforestation and degradation. Despite rapid improvements in Earth Observation technology, monitoring of tropical forests remains hindered by persistent cloud cover, heterogeneous landscapes, long wet seasons, and small and ephemeral clearings masked by rapid growth. A hybrid method is presented that combines elements of both time-series and compositing approaches to best overcome these obstacles to map forest cover and change in the Republic of Panama based on Landsat imagery. The resulting Panama Vegetation-Cover Time-Series (PVCTS) maps depict forest cover in Panama from 1990 to 2016 at 30 m resolution. Acknowledging the fuzzy boundary between forest and non-forest classes, these maps employ a hierarchical classification scheme that reflects the natural process of regeneration and can accommodate different definitions of forest and deforestation. Classification accuracy is 97-98 % between forest/non-forest categories and 76-81 % for deforestation events. The maps show a slight greening of Panama from 1990 to 2016 caused by expansion of young secondary growth. The annual rate of deforestation in mature forest has remained around-0.6 %/yr, although young forests have matured at a similar rate such that there is no net loss of forest. While estimates of total forest cover are similar to official national estimates depending on forest definition, there is little agreement in location of deforestation events.
... Indicators of spatio-temporal characteristics have increased attention on describing landscape changes . The indicators that have been widely used to describe such characteristics include change time, frequency, and rate of change, focusing on a single aspect of the ecological status with a single spectral index (Broich et al., 2011;Jong et al., 2012;Kennedy et al., 2010;Piao et al., 2015;Turubanova et al., 2015). Subsequently, some researchers have tried to develop spatial autocorrelation coefficients and spatial cluster analysis indices (Fan et al., 2017;Hermosilla et al., 2015). ...
... For example, recently published research assesses change in vegetation in terms of the growth or shrinkage of plant cover, but does not take into account the naturalness of the change [24,25]. Structural changes in vegetation have also been addressed in recent works that analyzed this parameter on the basis of its homogeneity [25] or some very broad vegetation classes [26]. ...
The goal of this study was to propose scientific and objective indices capable of measuring the changes that occur in the conservation status of the vegetation of a particular area over a period of time. To this end, phytosociologically-based vegetation cartography at a detailed scale was used, carried out at two different times, and the distance from the climax stage of the territory was calculated for each time. Three temporal indices of landscape change are proposed: Conservation Status Variation Index (ConSVI), Conservation Status Variation Velocity Index (ConSVVe) and Change Ratio (ChanRat). These enable the intensity, velocity, and percentage of change to be measured, and to determine whether this change is progressive or regressive-in other words, whether it is approaching or receding from the climax. To test the proposal, it was applied to a territory in Northwest Spain. The proposed indices are universally applicable to any territory and are the first of their kind to operate at a detailed scale with a phytosociological basis. They also enable an objective measurement to be made of the landscape change that has occurred, meaning that they have immense practical utility in studies of managing and planning territorial resources.
... We then calculated the minimum, maximum and selected percentile values (10, 25, 50, 75 and 90% percentiles) and the mean reflectance values for observations between selected percentiles (10-25%, 25-50%, 50-75%, 75-90%, and 25-75%). Similar time-series metrics have been successfully used in forest cover mapping using Landsat data (Broich et al., 2011;Potapov et al., 2012;Hansen et al., 2013). To further assist in differentiating between woody and herbaceous cover, which have different phenological metrics (Helman et al., 2015), we derived the variance and range in vegetation indices. ...
... The University of Maryland has performed time-series transformation of MODIS and Landsat data in Congo Basin [8], Landsat data from 1985 to 2012 in Eastern Europe's [9]. Image transformation is done by histogram-based metrics approach, such as average, and by sequential metric approach [8]- [10]. ...
This paper presents the paddy field classification model using the approach based on periodic plant life cycle events and how these elevations in climate as well as habitat factors, such as elevation. The data used are MODIS-Terra two tiles of H28v09 and H29v09 of 2016, consist of 46 series of 8-daily data, with 500 meter resolution in Java region. The paddy field classification method based on the phenological model is done by Maximum Likelihood on the transformed annual multi-temporal image of the reflectance data, index data, and the combination of reflectance and index data. The results of the study showed that, with the reference of the Paddy Field Map from the Ministry of Agriculture (MoA), the overall accuracies of the paddy field classification results using the combination of reflectance and index data provide the highest (85.4%) among the reflectance data (83.5%) and index data (81.7%). The accuracy levels were varied; these depend on the slope and the types of paddy fields. Paddy fields on the slopes of 0-2% could be well identified by MODIS-Terra data, whereas it was difficult to identify the paddy fields on the slope >2%. Rain-fed lowland paddy field type has a lower user accuracy than irrigated paddy fields. This study also performed correlation (r2) between the analysis results and the statistical data based on district and provincial boundaries were >0.85 and >0.99 respectively. These correlations were much higher than the previous study results, which reached 0.49-0.65 (hilly-flat areas of county-level), and 0.80-0.88 (hilly-flat areas of provincial level) for China, and reached 0.44 for Indonesia.
... Optical satellite imagery is susceptible to cloud cover, particularly in humid and tropical regions where cloud cover and optically dense atmospheres are predominant. Much research has demonstrated the potential of image composites, formed by combining multitemporal cloud-free observations, to resolve cloud cover and cloud shadow issues (Broich et al. 2011;Huang et al. 2009;Hansen et al. 2008). Lindquist et al. (2008) evaluated the pixel quality for the humid tropics in central Africa and found that, ideally, all available image data would be utilized to achieve the highest image composite quality. ...
Rapid population and economic growth quickly degrade and deplete forest resources in many developing countries, even within protected areas. Monitoring forest cover change is critical for assessing ecosystem changes and targeting conservation efforts. Yet the most biodiverse forests on the planet are also the most difficult to monitor remotely due to their frequent cloud cover. To begin to reconcile this problem, we develop and implement an effective and efficient approach to mapping forest loss in the extremely cloud-prevalent southern Ghana region using dense time series Landsat 7 and 8 images from 1999 to 2018, based on median value temporal compositing of a novel vegetation index called the spectral variability vegetation index (SVVI). Resultant land-cover and land-use maps yielded 90 to 94% mapping accuracies. Our results indicate 625 km² of forest loss within the 9800-km² total mapping area, including within forest reserves and their environs between circa 2003 and 2018. Within the reserves, reduced forest cover is found near the reserve boundaries compared with their interiors, suggesting a more degraded environment near the edge of the protected areas. A fully protected reserve, Kakum National Park, showed little forest cover change compared with many other less protected reserves (such as a production reserve—Subri River). Anthropogenic activities, such as mining, agriculture, and built area expansion, were the main land-use transitions from forest. The reserves and census districts that are located near large-scale open pit mining indicated the most drastic forest loss. No significant correlation was found between the magnitudes of forest cover change and population density change for reserves and within a 1.5-km buffer surrounding the reserves. While other anthropogenic factors should be explored in relation to deforestation, our qualitative analysis revealed that reserve protection status (management policies) appears to be an important factor. The mapping approach described in this study provided a highly accurate and effective means to monitor land-use changes in forested and cloud-prone regions with great promise for application to improved monitoring of moist tropical and other forests characterized by high cloud cover.
... Mapping these species using currently available satellite data involves decisions concerning the relative importance of spectral and spatial resolution (May et al., 1997). Changes in the spatial and spectral resolution help not only in providing meaningful land use/cover maps but also in monitoring natural resources and environmental degradation (Duong et al., 2000;Broich et al., 2011). ...
Monitoring tropical rain forests via remotely sensed imagery has become very useful in understanding land use/cover change over time for three East Africa rain forest areas: Kakamega-Nandi forests area in Kenya, Mabira and Budongo Forest areas in Uganda. Landsat imagery, provided free of charge by the U.S. Geological Survey Center for Earth Resources Observation and Science (USGS/EROS), enables consistent and timely forest cover change over time. Analyses of the landuse/ cover changes since the early 1970s until 2003 for these three rain forests were done by processing Landsat Multispectral Scanner (MSS), Thematic Mapper (TM) and Enhanced Thematic Mapper plus (ETM+) imagery for eight or seven time steps at regular intervals by Biodiversity Monitoring Transect Analysis in Eastern Africa (BIOTA East Africa). For continuous forest change analysis, the three existing time series data are to be extended by another two time steps from more recent years (2005/2006 and 2007/2008) as part of the remote sensing activities within BIOTA East Africa sub-project E02. Since, on 31 May 2003 L andsat ETM+ suffered the loss of its scan line corrector (SLC) which removes the "zigzag" motion of the imaging field of view produced by the combination of the along and cross track motion, there is data loss of a bout 22% of the total area of the scene, then there is need to get alternative solution. This study describes a methodology for combining several SLC-off images of Budongo Forest area into a one single dataset to be used as basis for land use/cover classification. The approach of filling gaps used in this method involved techniques of adding classified images in order to come up with a meaningful classification. Two images per time step are used to come up with one meaningful classification. Additionally, the suitability of SPOT-4 multispectral image data for deriving land use/cover classifications for another two time steps of Kakamega Nandi and Mabira forests areas have been investigated to give truly comparable results to the existing Landsat-derived time series data. Both SPOT and Landsat SLC-off data offered the chance of extending the existing times series with truly comparable classification results. The same land cover classes have been distinguished as in the previous time steps using supervised multispectral classification. The applied methodology resulted in high classification accuracies.
... Remote sensing has been frequently used to monitor changes in mangrove forest cover at local and global levels, e.g., [11][12][13][14][15][16][17][18][19][20][21][22][23]. With an increasing array and availability of satellite sensor data, there is an opportunity to routinely observe mangrove forests and support silvicultural management, restoration and reforestation projects. ...
Time series of satellite sensor data have been used to quantify mangrove cover changes at regional and global levels. Although mangrove forests have been monitored using remote sensing techniques, the use of time series to quantify the regeneration of these forests still remains limited. In this study, we focus on the Matang Mangrove Forest Reserve (MMFR) located in Peninsular Malaysia, which has been under silvicultural management since 1902 and provided the opportunity to investigate the use of Landsat annual time series (1988–2015) for (i) detecting clear-felling events that take place in the reserve as part of the local management, and (ii) tracing back and quantifying the early regeneration of mangrove forest patches after clear-felling. Clear-felling events were detected for each year using the Normalized Difference Moisture Index (NDMI) derived from single date (cloud-free) or multi-date composites of Landsat sensor data. From this series, we found that the average period for the NDMI to recover to values observed prior to the clear-felling event between 1988 and 2015 was 5.9 ± 2.7 years. The maps created in this study can be used to guide the replantation strategies, the clear-felling planning, and the management and monitoring activities of the MMFR.
... The University of Maryland has performed time-series transformation of MODIS and Landsat data in Congo Basin [8], Landsat data from 1985 to 2012 in Eastern Europe's [9]. Image transformation is done by histogram-based metrics approach, such as average, and by sequential metric approach [8]- [10]. ...
This paper presents the paddy field classification model using the approach based on periodic plant life cycle events and how these elevations in climate as well as habitat factors, such as elevation. The data used are MODIS-Terra two tiles of H28v09 and H29v09 of 2016, consist of 46 series of 8-daily data, with 500 meter resolution in Java region. The paddy field classification method based on the phenological model is done by Maximum Likelihood on the transformed annual multi-temporal image of the reflectance data, index data, and the combination of reflectance and index data. The results of the study showed that, with the reference of the Paddy Field Map from the Ministry of Agriculture (MoA), the overall accuracies of the paddy field classification results using the combination of reflectance and index data provide the highest (85.4%) among the reflectance data (83.5%) and index data (81.7%). The accuracy levels were varied; these depend on the slope and the types of paddy fields. Paddy fields on the slopes of 0-2% could be well identified by MODIS-Terra data, whereas it was difficult to identify the paddy fields on the slope >2%. Rain-fed lowland paddy field type has a lower user accuracy than irrigated paddy fields. This study also performed correlation (r2) between the analysis results and the statistical data based on district and provincial boundaries were >0.85 and >0.99 respectively. These correlations were much higher than the previous study results, which reached 0.49-0.65 (hilly-flat areas of county-level), and 0.80-0.88 (hilly-flat areas of provincial level) for China, and reached 0.44 for Indonesia. Copyright © 2019 Institute of Advanced Engineering and Science. All rights reserved.
... Hutan di Indonesia tersusun oleh beragam mosaik dengan kualitas lahan dan tegakan yang beragam bahkan ada yang sangat terdegradasi (Kusmana 2011). Luas hutan alam di Indonesia terus berkurang dari tahun ke tahun (Nursanti 2008;Hansen et al. 2009;Broich et al. 2011;KLHK 2015;KLHK 2016). Berdasarkan KLHK 2017 luas total kawasan hutan di Indonesia adalah 125.96 juta ha dan 32.70 juta ha dari luas kawasan hutan tersebut dalam kondisi terdegradasi. ...
Natural forest in Indonesia was divided into several conditions that become a
mosaic of forest areas. Deforestation and forest degradation were occurring in the
existing forests. Technical efforts are needed to improve the condition of damaged
forests. One of the improvement efforts is to use a new silviculture system that can
improve the existing forest conditions. Selective logging and gap planting (TPTR)
is expected to be a solution that can be applied in multisystem silviculture. The
objective of this study was to obtain information about the impacts of TPTR
silviculture systems to soil compactness, erosion and fire danger.
There were 16 gaps in this study that were divided into 8 classes. Gap area
was measured with sixteen-gon method. Soil compaction, erosion, temperature
different and light intensity were analyzed with ANOVA test.
Soil compaction was measured with cone penetrometer. Soil compaction
decrease based on forest conditions were, respectively, crop strip (3.38 kgf/cm2) >
slash strip (3.23 kgf/cm2) > natural forest (2.76 kgf/cm2). The average of soil
compaction classified as very loose soil condition. The ANOVA test showed that
there were significant difference of soil compaction in 1st and 6th gaps class.
Erosion decrease of forest conditions were, respectively, natural forest
(25.50 m3 ha-1 or 17.55 ton ha-1) > crop strip (17.44 m3 ha-1 or 11.91 ton ha-1) >
slash strip (6.58 ton ha-1 or 3.79 ton ha-1). Erosion was measured using a field
observation method. Topographic differences, forest conditions and rainfall
decrease-increase patterns were factors that influenced the eroded soil amount.
ANOVA test results showed not any significant result of erosion magnitude
between forest conditions.
The temperature and light intensity difference were measured with
thermometer and luxmeter LX-1010b. The temperature average inside the gaps
about 28.9°C to 33.3°C and outside gaps about 28.7°C to 29.9°C. Light intensity
that entered inside the gap about 8258 to 39 242 lux. The temperature and light
intensity inside the gap were suitable for Anthocephalus spp. growth. The highest
forest fire danger rating inside the gaps was 3rd level. Temperature and light
intensity were significant different inside the gap and outside the gap. In the other
hand, there was no forest fire danger significant different inside and outside the gap.
Based on ecological impact, the gap with area of 1750–2000 m2 was the
optimal gap size. The optimal gap size that can be created in the field based on total
utility values were about 250–1000 m2 and 1250–2000 m2.
The global extent of the amount of burned area seems to have changed substantially in the last two decades. Discussions regarding the main force behind the current trends have dominated research in recent years, with several studies attributing the global decline in wildfires to socio-economic and land-use changes. This review discusses the uncertainties and limitations of remotely sensed data used to determine global trends in burned areas and changes in their potential drivers. In particular, we quantify changes in the amount of burned area and cropland area and illustrate the lack of consistency in the direction and magnitude of the trend in cropland land cover type specifically within sub-Saharan Africa, the region where data show a strong trend in the amount of burned area. We state the limitations of remote-sensed fire and land cover products. We end by demonstrating that based on the currently available data and research methods applied in the literature, it is not possible to unequivocally determine that cropland expansion is the primary driver of the decline in fire activity.
We proposed a new image compositing algorithm (MAX-RNB) based on the maximum ratio of Near Infrared (NIR) to Blue band (RNB), and evaluated it together with nine other compositing algorithms: MAX-NDVI (maximum Normalized Difference Vegetation Index), MED-NIR (median NIR band), WELD (conterminous United States Web-Enabled Landsat Data), BAP (Best Available Pixel), PAC (Phenology Adaptive Composite), WPS (Weighted Parametric Scoring), MEDOID (medoid measurement), COSSIM (cosine similarity), and NLCD (National Land Cover Database). Each algorithm was applied to time series of Landsat observations collected within two separate years at six locations around the world, to produce monthly (July 1 ± 15 days), 2 seasonal (July 1 ± 45 days), and annual (July 1 ± 180 days) composite images free of cloud, cloud shadow, and snow/ice. By comparing the composite images to reference Landsat images acquired in the growing season (closest to July 1 within ± 15 days) for each year, we evaluated the performance of the algorithms in preserving the spectral and spatial fidelity (hereafter referred to as spectral and spatial evaluation, respectively), as well as land cover classification and land change detection (hereafter referred to as application evaluation). The results demonstrated that no single algorithm outperformed all other algorithms in all the evaluations, but that performance depended on compositing intervals and cloud cover. For monthly composites, the MAX-RNB algorithm generally produced the best results in the spectral and application evaluations. For seasonal composites, the NLCD algorithm produced the best results in the spectral and application evaluations. For annual composites, the PAC algorithm produced the best results in the spectral evaluation and change detection, whereas BAP produced the best results in land cover classification. The BAP algorithm also produced the best results in the spatial evaluation for all the compositing periods. This study provides a comprehensive guidance for selecting the most appropriate image compositing algorithm for different Landsat-based applications.
Precise large-area crop mapping products are the foundation for cropland monitoring, market policy decision-making, and subsequent agricultural applications. However, owing to sharp discrepancies among the temporal-spectral attributes of crops from different areas, crop mapping across phenological regions faces significant challenges. Given the inadequacy of generalizing mapping frameworks from the source to target regions, research areas are often limited, and trained models cannot be reused beyond a particular study area. In this paper, we propose a novel framework called the Phenology Alignment Network (PAN) to address the cross-phenological-region (CPR) crop-mapping problem using deep recurrent networks and unsupervised domain adaptation. Our PAN adopts a Siamese structure consisting of two identical deep models called the Temporal Spectral Network (TSNet), which serves as an adaptive multilevel phenological feature extractor for crops under various planting conditions. Specifically, the two branches of PAN accept crop samples from the source and target regions, and aim to encode them into similar deep phenological features. By constraining the distribution of hidden states from the two branches, the deep temporal-spectral features are aligned, and regional discrepancies between the source and target regions are alleviated. Consequently, the PAN adapts the deep model pretrained on the source region to the target area, improving the mapping accuracy without using additional target label information. To verify the effectiveness of our model, crop samples were collected from three plains of China, and a winter crop dataset based on Sentinel-2 time-series imagery containing over two million pixels was annotated. Experiments conducted on this dataset indicate that PAN serves as a practical CPR crop mapping framework and achieves significant improvement in the overall accuracy as well as the macro-average F1 score compared to other state-of-the-art methods. Further visual interpretation illustrates the capability of PAN to correct inaccurate predictions ascribed to phenological diversity and suggests the potential of PAN for large-area crop mapping applications.
Synthetic aperture radar (SAR) has been a powerful tool for deforestation detection in tropical rainforests. Polarimetric SAR (POLSAR) data are acquired in a quad-polarization mode, and L-band POLSAR data in particular are one of the few SAR data types that preserve the dielectric properties and structures of the scatterer. POLSAR data consist of 2 × 2 scattering matrices and consequently offer superior target recognition compared with dual-polarization data. In this study, we applied scattering power decomposition suitable for detecting deforestation in near real-time to POLSAR data obtained from the Earth observation sensor Phased Array type L-band Synthetic Aperture Radar-2 (PALSAR-2). The reflection symmetry condition is known to apply to natural distributed objects (i.e., the cross-correlation between co- and cross-polarization data is zero). Inspired by this, we theoretically and experimentally examined the volume scattering power component to distinguish natural forests from the surrounding area. Two important results were verified for natural forests: as the window size of the ensemble average increases, (i) the coherency matrix approaches a simple theoretical form and (ii) the volume scattering power becomes dominant among the scattering power components. Based on these results, we constructed an algorithm that applies scattering power decomposition for detecting deforestation. We produced reference data using high-resolution optical images and evaluated the performance of the derived deforestation map in the Amazon natural forest when employing various window sizes for the ensemble average. At an optimal window size of 15 × 15 pixels, the deforestation detection performance reached a user's accuracy of 94.9% ± 1.5%, a producer's accuracy of 72.3% ± 1.2%, and a kappa coefficient of 0.816 ± 0.0039. Sparse trees left after logging increased the volume scattering power and reduced the producer's accuracy. The proposed algorithm can contribute to deforestation detection with slightly lower accuracy than that of the annual map provided by Global Forest Change. Further, the proposed algorithm is robust to the seasonal variations in tropical rainforests and temporal variations in the deforestation process. Consequently, the proposed algorithm employing the six-component scattering power decomposition method can be utilized in near real-time without considering applicable areas in tropical rainforests. Subsequently, we applied our algorithm to dual-polarization data, which are acquired by PALSAR-2 much more frequently than POLSAR data. The false detection rate did not increase when using the dual-polarization data; however, the omission error increased considerably compared with that when using POLSAR data owing to the low total power obtained from the dual-polarization data.
This article explores a potential exploitation of fully polarimetric radar data for the management of rubber plantations, specifically for predicting tree circumference as a crucial information need for sustainable plantation management. Conventional backscatter coefficients along with Eigen-based and model-based decomposition features served as the predictors in models of tree girth using ten regression approaches. The findings suggest that backscatter coefficients and Eigen-based decomposition features yielded lower accuracy than model-based decomposition features. Model-based decompositions, especially the Singh decomposition, provided the best accuracies when they were coupled with guided regularized random forests regression. This research demonstrates that L-band SAR data can provide an accurate estimation of rubber plantation tree girth, with an RMSE of about 8 cm.
Indonesia's natural forests consist of various land and stand qualities, both highly degraded. In the past few years, the forest area has continued to decline. Gap planting is one of the silviculture techniques which can increase the productivity of the low potential tropical natural forest, mainly by using fast-growing species with short cutting rotation. The objective of this research was to study the impacts of gap planting on soil compaction and erosion in the Indonesian lowland tropical rain forest. Soil compaction was measured using Humbolt-Digital Statis Cone Penetrometer, soil erosion was measured using erosion pins and gap area was measured using Hexadecagon Method. The result shows that using gap planting causes soil compaction in gaps with an area of less than or equal to 250 m 2 and 1250-1500 m 2. However, the compaction value is classified as very loose soil so that it can be ignored. Otherwise, gap planting, using planting strips and chop off strips, has a positive impact by lowering erosion at gap size less than or equal to 250 to 2000 m 2 .
Despite substantial research conducted within the forestry domain, detailed assessments to monitor plantations and support their sustainable management have been understudied. This article attempts to fill this gap through coupling fully polarimetric L-band data and contemporary data mining methods for the estimation of tree circumference as: (1) a primary dataset for biomass accumulation studies; and, (2) critical information for operational management in rubber plantations. We used two rubber plantation sites in Subang (West Java) and Jember (East Java), Indonesia, to evaluate the capability of L-band radar data. Although polarimetric features derived from polarimetric decomposition theorems have been advocated by others, we show that backscatter coefficients, especially HV polarization, remain an important dataset for this research domain. Using Subang data to build the model, we found that modern machine learning methods do not always deliver the best performance. It appears that the data being ingested plays a significant role in obtaining a good model, hence careful selection of datasets from multiple forms of polarimetric SAR data needs to be further considered. The highest coefficient of determination (R² = 0.79) was achieved by Yamaguchi decomposition features with the aid of partial least squares regression. Nonetheless, we note that the R² gap was insignificant to the backscatter coefficient when random forests regression was used (R² = 0.78). Overall, only the backscatter coefficient dataset delivered fairly consistent results with any regression model, with the average R² being about 0.67. When tuning parameters were not assessed, random forests consistently outweighed support vector regressions in all forms of datasets. The latter generated a substantial increase in R² when a linear kernel was used instead of the popular radial basis function. The issue of transferability of the model is also addressed in this article. It appears that similarity of terrain characteristics substantially influences the model’s performance. Models developed in Subang, which has gentle slopes, seem valid only in plantations with similar terrain. Validation attempts in very flat terrain within two plantation sectors in Jember delivered a poor result, although they have similar elevations to the Subang site. In contrast, validation in a plantation sector with similar, gently sloping terrain achieved an R² of about 0.6 using some datasets.
We assess the magnitude and the extent of recent change of significant human footprint within protected areas, key biodiversity areas and the habitat range of 308 lowland forest specialist birds in Sundaland, a global hotspot of biodiversity in Southeast Asia. Using the most recent human footprint dataset, we find that 70% of Sundaland has been heavily modified by humans. This represents a 55% increase in areas under intense human pressure since 1993. Areas under intense human pressure covered on average 50% of the extent of key biodiversity areas, 78% of each protected area and 38% of the range of lowland forest specialist birds. The results imply that the actual level of protection by protected areas is only one‐third to half of that on paper once human footprint is accounted for. While all protected areas were impacted by human pressures, those managed strictly for biodiversity conservation presented the largest increases. These results highlight an exceptionally high human footprint across Sundaland and an impending further deepening of the biodiversity crisis across the region.
This study develops a modelling framework by utilizing multi-sensor imagery for classifying different forest and land use types in the Phnom Kulen National Park (PKNP) in Cambodia. Three remote sensing datasets (Landsat optical data, ALOS L-band data and LiDAR derived Canopy Height Model (CHM)) were used in conjunction with three different machine learning (ML) regression techniques (Support Vector Machines (SVM), Random Forests (RF) and Artificial Neural Networks (ANN)). These ML methods were implemented on (a) Landsat spectral data, (b) Landsat spectral band & ALOS backscatter data, and (c) Landsat spectral band, ALOS backscatter data, & LiDAR CHM data. The Landsat-ALOS combination produced more accurate classification results (95% overall accuracy with SVM) compared to Landsat-only bands for all ML models. Inclusion of LiDAR CHM (which is a proxy for vertical canopy heights) improved the overall accuracy to 98%. The research establishes that majority of PKNP is dominated by cashew plantations and the nearly intact forests are concentrated in the more inaccessible parts of the park. The findings demonstrate how different RS datasets can be used in conjunction with different ML models to map forests that had undergone varying levels of degradation and plantations.
Information needs associated with forest monitoring have become increasingly complex. Data to support these information needs are required to be systematically generated, spatially exhaustive, spatially explicit, and to capture changes at a spatial and temporal resolution that is commensurate with both natural and anthropogenic impacts. Moreover, reporting obligations impose additional expectations of transparency, repeatability, and data provenance. The overall objective of this dissertation was to address these needs and improve capacity for large-area monitoring of forest disturbance and subsequent recovery. Landsat time series (LTS) enhance opportunities for forest monitoring, particularly for post-disturbance recovery assessments, while best-available pixel (BAP) compositing approaches allow LTS approaches to be applied over large forest extents. In substudies I and IV, forest monitoring information needs were identified and linked to image compositing criteria and data availability in Canada and Finland. In substudy II, methods were developed and demonstrated for generating large-area, gap-filled Landsat BAP image composites that preserve detected changes, generate continuous change metrics, and provide foundational, annual data to support forest monitoring. In substudy III a national monitoring framework was prototyped at scale over the 650 Mha of Canada’s forest ecosystems, providing a detailed analysis of areas disturbed by wildfire and harvest for a 25-year period (1985–2010), as well as characterizing short- and long-term recovery. New insights on spectral recovery metrics were provided by substudies V and VI. In substudies V, the utility of spectral measures of recovery were evaluated and confirmed against benchmarks of forest cover and height derived from airborne laser scanning data. In substudy VI the influence of field-measured structure and composition on spectral recovery were examined and quantified. By focusing on four key aspects of forest monitoring systems: information needs, data availability, methods development, and information outcomes, the component studies demonstrated that combining BAP compositing and LTS analysis approaches provides data with the requisite characteristics to support large-area forest monitoring, while also enabling a more comprehensive assessment of forest disturbance and recovery.
In this paper three methods for updating inventories of burned areas have been presented and examined. They include Multitemporal Principal Component Analysis (MPCA), Change Vector Analysis (CVA) and Multitemporal NDVI Classification (MNC). First, 11 Landsat-5 Thematic Mapper (TM) images of a forest area were radiometrically corrected to derive a multitemporal series of intercomparable images for each spring from 1984 to 1994. Then, in order to check the feasibility of the three approaches, they were used for mapping fire burns that occurred during 1992. The various procedures yielded different maps of burned areas; the MNC method seemed to be more reliable than the others, because it merges spectral data corresponding not only to 1992 (pre-fire) and 1993 (post-fire) but also to 1994 (the second year after the fires), which is key in the vegetation regeneration. Finally, this methodology was automated to yield an inventory of burned areas for each year during the period of study.
Landsat remote sensing of the central African humid tropics is confounded by persistent cloud cover and, since 2003, missing data due to the Landsat‐7 Enhanced Thematic Mapper Plus (ETM+) scan line corrector (SLC) malfunction. To quantify these limitations and their effects on contemporary forest cover and change characterization, a comparison was made of multiple Landsat‐7 image mosaics generated for a six Landsat path/row study site in central Africa for 2000 and 2005. Epoch 2000 mosaics were generated by compositing (i) two to three Landsat acquisitions per path/row, (ii) using the best single GeoCover 2000 acquisition for each path/row. Epoch 2005 composites were generated by compositing SLC‐off data using (iii) five to seven acquisitions per path/row, (iv) three acquisitions per path/row. Eighty per cent of pixels were of suitable quality for change detection between (ii) and (iv), emulating that which is possible with current GeoCover and planned Global Land Survey (GLS) inputs. In a more data intensive change detection analysis using mosaics (i) and (iii), 96% of pixels had suitable quality. Compositing more acquisitions per path/row for the study area systematically reduced the percentage of SLC‐off gaps and, when more than three acquisitions were composited, reduced the percentage of pixels with high likelihood of cloud, haze or shadow. The results indicate that additional input imagery to augment both the Geocover and GLS data may be required to enable forest cover and change analyses for regions of the humid tropics.
The new SPOT/Envisat direct receiving station (DRS) operating in Cayenne in the framework of the SEAS-Guyane project was used to produce a global cloudless SPOT mosaic over French Guiana for the year 2006. This mosaic was used to perform a land-use, land-use change and forestry (LULUCF) inventory in the framework of the Kyoto Protocol. Nearly 17 000 sample points were laid down on the SPOT mosaic with a stratified sampling design. The land use at each sample point was determined by visual interpretation of the corresponding SPOT image in 2006 and the Landsat Thematic Mapper (TM) image from the Global Land Cover Facility (GLCF) in 1990. Statistics for the period 1990–2006 were computed and integrated in the first voluntary Kyoto inventory.
Techniques based on multi-temporal, multi-spectral, satellite-sensor- acquired data have demonstrated potential as a means to detect, identify, map and monitor ecosystem changes, irrespective of their causal agents. This review paper, which summarizes the methods and the results of digital change detection in the optical/infrared domain, has as its primary objective a synthesis of the state of the art today. It approaches,digital change,detection from,three angles. First, the different perspectives from which the variability in ecosystems and the change,events have been dealt with are summarized.,Change,detection between pairs of images,(bi-temporal) as well as between,time profiles of imagery,derived indicators (temporal trajectories), and, where relevant, the appropriate choices for digital imagery acquisition timing and change interval length definition, are discussed. Second, pre-processing routines either to establish a more direct linkage between remote sensing data and biophysical phenomena, or to temporally mosaic imagery and extract time profiles, are reviewed. Third, the actual change,detection,methods,themselves,are categorized,in an analytical framework and critically evaluated. Ultimately, the paper highlights how some of these methodological,aspects are being,fine-tuned as this review,is being written, and we summarize the new developments that can be expected in the near future. The review,highlights the high complementarity,between,different change,detection methods.
Six change detection procedures were tested using Landsat Multi-Spectral Scanner (MSS) images for detecting areas of changes in the region of the Terminos Lagoon, a coastal zone of the State of Campeche, Mexico. The change detection techniques considered were image differencing, vegetative index differencing, selective principal components analysis (SPCA), direct multi-date unsupervised classification, post-classification change differencing and a combination of image enhancement and post-classification comparison. The accuracy of the results obtained by each technique was evaluated by comparison with aerial photographs through Kappa coefficient calculation. Post-classification comparison was found to be the most accurate procedure and presented the advantage of indicating the nature of the changes. Poor performances obtained by image enhancement procedures were attributed to the spectral variation due to differences in soil moisture and in vegetation phenology between both scenes. Methods based on classification were found to be less sensitive at these spectral variations and more robust when dealing with data captured at different times of the year.
The first results of the Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation continuous field algorithm's global percent tree cover are presented. Percent tree cover per 500-m MODIS pixel is estimated using a supervised regression tree algorithm. Data derived from the MODIS visible bands contribute the most to discriminating tree cover. The results show that MODIS data yield greater spatial detail in the characterization of tree cover compared to past efforts using AVHRR data. This finer-scale depiction should allow for using successive tree cover maps in change detection studies at the global scale. Initial validation efforts show a reasonable relationship between the MODIS-estimated tree cover and tree cover from validation sites.
Timely and accurate change detection of Earth’s surface features is
extremely important for understanding relationships and interactions between
human and natural phenomena in order to promote better decision making.
Remote sensing data are primary sources extensively used for change detection
in recent decades. Many change detection techniques have been developed. This
paper summarizes and reviews these techniques. Previous literature has shown
that image differencing, principal component analysis and post-classification
comparison are the most common methods used for change detection. In recent
years, spectral mixture analysis, artificial neural networks and integration of
geographical information system and remote sensing data have become
important techniques for change detection applications. Different change
detection algorithms have their own merits and no single approach is optimal
and applicable to all cases. In practice, different algorithms are often compared
to find the best change detection results for a specific application. Research of
change detection techniques is still an active topic and new techniques are
needed to effectively use the increasingly diverse and complex remotely sensed
data available or projected to be soon available from satellite and airborne
sensors. This paper is a comprehensive exploration of all the major change
detection approaches implemented as found in the literature.
The world's forest ecosystems are in a state of permanent flux at a variety of spatial and temporal scales. Monitoring techniques based on multispectral satellite‐acquired data have demonstrated potential as a means to detect, identify, and map changes in forest cover. This paper, which reviews the methods and the results of digital change detection primarily in temperate forest ecosystems, has two major components. First, the different perspectives from which the variability in the change event has been approached are summarized, and the appropriate choice of digital imagery acquisition dates and interval length for change detection are discussed. In the second part, preprocessing routines to establish a more direct linkage between digital remote sensing data and biophysical phenomena, and the actual change detection methods themselves are reviewed and critically assessed. A case study in temperate forests (north‐central U.S.A.) then serves as an illustration of how the different change detection phases discussed in this paper can be integrated into an efficient and successful monitoring technique. Lastly, new developments in digital change detection such as the use of radar imagery and knowledge‐based expert systems are highlighted.
Timely and accurate data on forest change within Indonesia is required to provide government, private and civil society interests with the information needed to improve forest management. The forest clearing rate in Indonesia is among the highest reported by the United Nations Food and Agriculture Organization (FAO), behind only Brazil in terms of forest area lost. While the rate of forest loss reported by FAO was constant from 1990 through 2005 (1.87 Mha yr−1), the political, economic, social and environmental drivers of forest clearing changed at the close of the last century. We employed a consistent methodology and data source to quantify forest clearing from 1990 to 2000 and from 2000 to 2005. Results show a dramatic reduction in clearing from a 1990s average of 1.78 Mha yr−1 to an average of 0.71 Mha yr−1 from 2000 to 2005. However, annual forest cover loss indicator maps reveal a near-monotonic increase in clearing from a low in 2000 to a high in 2005. Results illustrate a dramatic downturn in forest clearing at the turn of the century followed by a steady resurgence thereafter to levels estimated to exceed 1 Mha yr−1 by 2005. The lowlands of Sumatra and Kalimantan were the site of more than 70% of total forest clearing within Indonesia for both epochs; over 40% of the lowland forests of these island groups were cleared from 1990 to 2005. The method employed enables the derivation of internally consistent, national-scale changes in the rates of forest clearing, results that can inform carbon accounting programs such as the Reducing Emissions from Deforestation and Forest Degradation in Developing Countries (REDD) initiative.
Payments for reduced carbon emissions from deforestation (RED) are now attracting attention as a way to halt tropical deforestation. Northern Sumatra comprises an area of 65 000 km² that is both the site of Indonesia's first planned RED initiative, and the stronghold of 92% of remaining Sumatran orangutans. Under current plans, this RED initiative will be implemented in a defined geographic area, essentially a newly established, 7500 km² protected area (PA) comprising mostly upland forest, where guards will be recruited to enforce forest protection. Meanwhile, new roads are currently under construction, while companies are converting lowland forests into oil palm plantations. This case study predicts the effectiveness of RED in reducing deforestation and conserving orangutans for two distinct scenarios: the current plan of implementing RED within the specific boundary of a new upland PA, and an alternative scenario of implementing RED across landscapes outside PAs.
Our satellite-based spatially explicit deforestation model predicts that 1313 km² of forest would be saved from deforestation by 2030, while forest cover present in 2006 would shrink by 22% (7913 km²) across landscapes outside PAs if RED were only to be implemented in the upland PA. Meanwhile, orangutan habitat would reduce by 16% (1137 km²), resulting in the conservative loss of 1384 orangutans, or 25% of the current total population with or without RED intervention. By contrast, an estimated 7824 km² of forest could be saved from deforestation, with maximum benefit for orangutan conservation, if RED were to be implemented across all remaining forest landscapes outside PAs. Here, RED payments would compensate land users for their opportunity costs in not converting unprotected forests into oil palm, while the construction of new roads to service the marketing of oil palm would be halted.
Our predictions suggest that Indonesia's first RED initiative in an upland PA may not significantly reduce deforestation in northern Sumatra and would have little impact on orangutan conservation because a large amount of forest inside the project area is protected de facto by being inaccessible, while lowland forests will remain exposed to the combined expansion of high-revenue plantations and road networks. In contrast, RED would be more effective in terms of its conservation impact if payments were extended to all remaining carbon-rich tropical forests, including lowland peat swamp forests, the preferred habitat for dense populations of orangutans, and if the construction of new roads was halted.
This paper presents a semi-automatic methodology for re scars map-ping from a long time series of remote sensing data. Approximately, a hundred MSS images from di erent Landsat satellites were employed over an area of 32 100 km 2 in the north-east of the Iberian Peninsula. The analysed period was from 1975 to 1993. Results are a map series of re history and frequencies. Omission errors are 23% for burned areas greater than 200 ha while commission errors are 8% for areas greater than 50 ha. Subsequent work based on the resultant re scars will also help in describing re regime and in monitoring post-re regeneration dynamics.
Aim This study determines whether the establishment of tropical protected areas (PAs) has led to a reduction in deforestation within their boundaries or whether deforestation has been displaced to adjacent unprotected areas: a process termed neighbourhood leakage.
Location Sumatra, Indonesia.
Methods We processed and analysed 98 corresponding LANDSAT satellite images with a c. 800 m2 resolution to map deforestation from 1990 to 2000 across 440,000 km2 on the main island of Sumatra and the smaller island of Siberut. We compared deforestation rates across three categories of land: (1) within PAs; (2) in adjacent unprotected land lying with 10 km of PA boundaries; and (3) within the wider unprotected landscape. We used the statistical method of propensity score matching to predict the deforestation that would have been observed had there been no PAs and to control for the generally remote locations in which Sumatran PAs were established.
Results During the period 1990–2000 deforestation rates were found to be lower inside PAs than in adjacent unprotected areas or in the wider landscape. Deforestation rates were also found to be lower in adjacent unprotected areas than in the wider landscape.
Main conclusions Sumatran PAs have lower deforestation rates than unprotected areas. Furthermore, a reduction in deforestation rates inside Sumatran PAs has promoted protection, rather than deforestation, in adjacent unprotected land lying within 10 km of PA boundaries. Despite this positive evaluation, deforestation and logging have not halted within the boundaries of Sumatran PAs. Therefore the long-term viability of Sumatran forests remains open to question.
In western Oregon, forest ecosystem processes are greatly affected by patterns of stand replacement disturbance. A spatially
explicit characterization of clear-cut logging and wildfire is a prerequisite to understanding the causes and consequences
of disturbance in this region. We analyzed stand replacement disturbance over 4.6 million forested hectares of three major
provinces in western Oregon between 1972 and 1995, contrasting the relative amounts of wildfire and harvest in each province
and comparing harvest statistics among the dominant land ownership categories. Clear-cut harvest and wildfire occurred over
19.9% and 0.7% of the study area, respectively. The moist Coast Range Province (CRP) was dominated by private industrial (PI)
ownerships and had the greatest concentration of cutting. The relatively dry Klamath Mountains Province (KMP) and the climatically
moderate Western Cascades Province (WCP) were dominated by public landowners and had lower concentrations of cutting and larger
amounts of wildfire than the CRP. Rates of harvest over time generally followed similar trends across landowners; it was lowest
in the early 1970s, peaked in the late 1980s and early 1990s, and then decreased to near 1970s levels by the mid-1990s. PI
landowners had harvest rates that were about two and one-half times as high as public owners throughout the study period.
Public and private nonindustrial (PNI) owners tended to have relatively small cutting units that remained spatially dispersed
throughout the study period, whereas the PI owners had larger individual cutting units that tended toward spatial aggregation
over time. Comparing the managed disturbance regimes with historical wild disturbance regimes can help us to understand the
relative impact of management regimes on ecosystems.
Satellite sensors are well suited to monitoring changes on the Earth's surface through provision of consistent and repeatable measurements at a spatial scale appropriate for many processes causing change on the land surface. Here, we describe and test a new conceptual approach to change detection of forests using a dense temporal stack of Landsat Thematic Mapper (TM) imagery. The central premise of the method is the recognition that many phenomena associated with changes in land cover have distinctive temporal progressions both before and after the change event, and that these lead to characteristic temporal signatures in spectral space. Rather than search for single change events between two dates of imagery, we instead search for these idealized signatures in the entire temporal trajectory of spectral values. This trajectory-based change detection is automated, requires no screening of non-forest area, and requires no metric-specific threshold development. Moreover, the method simultaneously provides estimates of discontinuous phenomena (disturbance date and intensity) as well as continuous phenomena (post-disturbance regeneration). We applied the method to a stack of 18 Landsat TM images for the 20-year period from 1984 to 2004. When compared with direct interpreter delineation of disturbance events, the automated method accurately labeled year of disturbance with 90% overall accuracy in clear-cuts and with 77% accuracy in partial-cuts (thinnings). The primary source of error in the method was misregistration of images in the stack, suggesting that higher accuracies are possible with better registration.
Sugar maple (Acer Saccharum Marsh.) damage resulting from a severe ice storm was modeled and mapped over eastern Ontario using pre- and post-storm Landsat 5 imagery and environmental data. Visual damage estimates in 104 plots and corresponding reflectance and environmental data were divided into multiple, mutually exclusive training and reference datasets for damage classification evaluation. Damage classification accuracy was compared among four methods: multiple regression, linear discriminant analysis, maximum likelihood, and neural networks. Using the best classifier, various stratification methods were assessed for potential inflationary effects on classification accuracy due to spatial proximity between training and reference data. Of the classifiers that were evaluated, neural networks performed best. Neural networks ‘learn’ training data accurately (94% overall), but classify proximate reference data less accurately (65%), and distant, spatially independent reference data least accurately (55%). Results indicate that, while remotely sensed and environmental data cannot discriminate among many levels of deciduous ice storm damage, they can by considered useful for differentiating areas of low to medium damage from areas of severe damage (69% accuracy). Such classification methods can provide regional damage maps more objectively than point-based visual estimates or aerial sketch mapping and aid in identification of areas of severe damage where management intervention may be advantageous.
Since January 2008, the U.S. Department of Interior / U.S. Geological Survey have been providing free terrain-corrected (Level 1T) Landsat Enhanced Thematic Mapper Plus (ETM+) data via the Internet, currently for acquisitions with less than 40% cloud cover. With this rich dataset, temporally composited, mosaics of the conterminous United States (CONUS) were generated on a monthly, seasonal, and annual basis using 6521 ETM+ acquisitions from December 2007 to November 2008. The composited mosaics are designed to provide consistent Landsat data that can be used to derive land cover and geo-physical and bio-physical products for detailed regional assessments of land-cover dynamics and to study Earth system functioning. The data layers in the composited mosaics are defined at 30 m and include top of atmosphere (TOA) reflectance, TOA brightness temperature, TOA normalized difference vegetation index (NDVI), the date each composited pixel was acquired on, per-band radiometric saturation status, cloud mask values, and the number of acquisitions considered in the compositing period. Reduced spatial resolution browse imagery, and top of atmosphere 30 m reflectance time series extracted from the monthly composites, capture the expected land surface phenological change, and illustrate the potential of the composited mosaic data for terrestrial monitoring at high spatial resolution. The composited mosaics are available in 501 tiles of 5000 × 5000 30 m pixels in the Albers equal area projection and are downloadable at http://landsat.usgs.gov/WELD.php. The research described in this paper demonstrates the potential of Landsat data processing to provide a consistent, long-term, large-area, data record.
The long-term acquisition plan (LTAP) was developed to fulfill the Landsat-7 (L7) mission of building and seasonally refreshing an archive of global, essentially cloud-free, sunlit, land scenes. The LTAP is considered one of the primary successes of the mission. By incorporating
seasonality and cloud avoidance into the decision making used to schedule image acquisitions, the L7 data in the U.S. Landsat archive is more complete and of higher quality than has ever been previously achieved in the Landsat program.
Development of the LTAP system evolved over
more than a decade, starting in 1995. From 2002 to 2004 most attention has been given to validation of LTAP elements. We find that the original expectations and goals for the LTAP were surpassed for Landsat 7. When the L7 scan line corrector mirror failed, we adjusted the LTAP operations,
effectively demonstrating the flexibility of the LTAP concept to address unanticipated needs. During validation, we also identified some seasonal and geographic acquisition shortcomings of the implementation: including how the spectral vegetation index measurements were used and regional/seasonal
cloud climatology concerns. Some of these issues have already been at least partially addressed in the L7 LTAP, while others will wait further attention in the development of the LTAP for the Landsat Data Continuity Mission (LDCM). The lessons learned from a decade of work on the L7 LTAP provide
a solid foundation upon which to build future mission LTAPs including the LDCM.
Bagging predictors is a method for generating multiple versions of a predictor and using these to get an aggregated predictor. The aggregation averages over the versions when predicting a numerical outcome and does a plurality vote when predicting a class. The multiple versions are formed by making bootstrap replicates of the learning set and using these as new learning sets. Tests on real and simulated data sets using classification and regression trees and subset selection in linear regression show that bagging can give substantial gains in accuracy. The vital element is the instability of the prediction method. If perturbing the learning set can cause significant changes in the predictor constructed, then bagging can improve accuracy.
The concept of mixed pixels allows the interpretation of remote sensing digital image data at sub-pixel level. Fraction-image data, obtained using the notion of mixed pixels, offer a potentially powerful method to detect changes in land-cover over a given period of time. This study proposes a new approach to detect land-cover changes, using two sets of fraction-image data obtained from sets of multispectral image data acquired at two different dates, over the same area. Changes based on the selected pixel components are then used to generate the fraction-change image data, including both positive (increase) and negative (decrease) changes in each component. The proposed analysis is then performed in the fraction-change space in two different ways: (1) by implementing unsupervised classification methods and (2) by comparing the fraction-change images among themselves. The proposed methodology is tested on two sets of Landsat Thematic Mapper (TM) multispectral image data obtained at two different dates and covering a test area mapped in previous works. Results obtained by the proposed methodology are presented and discussed.