Interpretation Of Satellite Imagery Research Articles

Tracking Deforestation with Neural Networks

Forests are key to Earth’s biodiversity and the global ecosystem, hosting 80% of the planet’s plant biomass and two-thirds of all mammal species. However, they are also in decline; each year, around five million hectares of woodland are lost to logging and wildfires. Tracking this loss is essential to controlling and mitigating deforestation, but this requires careful interpretation of satellite imagery. Yiqiao Yin, Kieran Pichai, Benjamin Park and Aaron Park have developed a model to automatically identify forests from these images. Their new approach is already showing huge promise.

Integrated Workflow for Interpretation of Satellite Imageries Using Machine Learning to Predict Algal Bloom in Lake

Integrated Workflow for Interpretation of Satellite Imageries Using Machine Learning to Predict Algal Bloom in Lake

Incorporating interpreter variability into estimation of the total variance of land cover area estimates under simple random sampling

Area estimates of land cover and land cover change are often based on reference class labels determined by analysts interpreting satellite imagery and aerial photography. Different interpreters may assign different reference class labels to the same sample unit. This interpreter variability is typically not accounted for in variance estimators applied to area estimates of land cover. A simple measurement model provides the basis for an estimator of the total variance (VTotal) that takes into account both sampling variance and interpreter variance. This method requires two or more reference class interpretations (i.e., repeated measurements) obtained by analysts, working independently of each other, for the full sample or a random subsample of the full sample. Estimators of the total variance (V̂Total) and the variance component attributable to interpreters (V̂1) were obtained for the case of two reference class interpretations per repeated sample unit. To evaluate the effect of interpreter variability on variance estimation, we used land cover reference data interpreted by seven analysts who each interpreted the same 300 sample pixels from a region of the Pacific Northwest of the United States. From these data, we estimated the contribution of interpreter variance to the total variance (i.e., V̂1/V̂Total) and the relative bias of the standard simple random sampling variance estimator (V̂stand) as an estimator of VTotal, defined as 100%*(V̂stand−V̂Total)/V̂Total. For each of five land cover classes, we computed V̂1, V̂Total, and V̂stand using the sample data from each of the 21 possible pairwise combinations of the seven interpreters, and then calculated the mean of V̂1/V̂Total and the mean of the estimated relative bias of V̂stand over these 21 pairs. Based on the mean of V̂1/V̂Total per class, interpreter variance contributed from 25% (cropland) to 76% (grass/shrub) of the total variance, indicating that interpreter variance was a non-negligible component of the total variance. Typically, the standard variance estimator, V̂stand, underestimated the total variance with the mean estimated relative bias ranging from −3% (cropland) to −33% (grass/shrub). Classes with greater inconsistency between pairs of interpreters had larger contributions of interpreter variance to the total variance (V̂1/V̂Total) and larger negative estimated relative bias of V̂stand. Given that interpreter variance can contribute substantially to the total variance, the repeated measurements approach offers a practical way to incorporate this variability into an estimator of the total variance.

Identification of Domain Complexes in the Danube Delta Coastal Belt Using the Biophysical Cross-shore Classification System (BCCS), Based on Interpretation of Satellite Imagery

Finkl, C.W. and Makowski, C., 2022. Identification of domain complexes in the Danube Delta Coastal Belt using the Biophysical Cross-shore Classification System (BCCS), based on interpretation of satellite imagery. Journal of Coastal Research, 38(1), 1–18. Coconut Creek (Florida), ISSN 0749-0208. The Biophysical Cross-shore Classification System (BCCS) was applied to the Danube Delta on the NW margin of the Black Sea to ascertain whether this method of classification was applicable to extensive low-lying and low-relief wetland coastalscapes. For organizational purposes, the coastal belt study area was broken into four previously recognized delta lobes (Kilia, Sulina, St. George II, and Cosna–Sinoie) and the Chituc Formation (beach ridge plain) because these provided distinctive coastal segments that were manageable for cross-shore classification. The village of Prymors'ke in the Lyman municipality of Ukraine and the onshore Chituc Formation beach ridge strandplain on the southern margin of the delta marked the northern- and southernmost extents of the Danubian classification effort, respectively. Codified cross-shore catenary sequences were based on the occurrence of recurring archetypes (i.e. Barrier, Beach, Beach Ridge, Delta, Developed, Lagoon, and Wetland) that had alongshore extent and could be identified as domains. Sub archetypes were also interpreted and added to the cross-shore sequences to further refine the Danube Delta classification. Because long cross-shore catenary sequences included inland biophysical and anthropogenic units that often differed dramatically from immediate onshore features, the conflated catenas were bifurcated into onshore (seaward) and inland (landward) domains, sections, or branches that were collectively referred to as domain complexes. Extension of cross-shore transects up to 20 km inland to incorporate all salient aspects of the delta provided useful subdivision of delta landscapes. Sixteen cross-shore transects were used to derive 12 domain complexes extending 167.5 km along the Danubian coastal belt that reached from the approximate latitudes of Prymors'ke, Ukraine, in the north to about Corbu in Romania to the south. Classification of the Danube Delta coastal belts showed that it is possible to apply the BCCS to large-area delta regions using extended cross-shore transects with bifurcated catenary sequences that together describe seaward (onshore) and landward (inland) sections of the delta in the form of domain complexes.

Analysis on the Evolution of Rural Settlement Pattern and Its Influencing Factors in China from 1995 to 2015

Since the early 1990s, China has experienced rapid industrialization and urbanization. As cities have expanded rapidly, the spatial patterns of rural settlements also changed significantly. This study uses land use data from satellite imagery interpretation, socioeconomic statistics, and field survey data, together with techniques including landscape pattern analysis, kernel density estimation, and spatial measurement models, to analyze the evolving spatial patterns of rural settlements influencing factors in China from 1995 to 2015. The results indicate the following: First, China’s rural settlements experienced significant changes in the period 1995–2015, as 88.92% of Prefectural-level administrative district units saw an increase in rural settlement area, with total settlement size increasing by 1.35 million hectares, and settlement area sprawl index values can be summarized as “high in the west and low in the east”. Second, in the two-decade study period, the population agglomeration capacity of rural settlements in China continuously weakened, and the shape and structure of rural settlement became more complex and irregular. The scale and scope of the disappearance of rural settlement areas in the northeast and southeast regions was relatively drastic, and the kernel density value of settlements dropped significantly. Third, the increase in rural settlement land area is concentrated in low-altitude and low-slope areas, with a significant tendency to be near water and roads. Fourth, social and economic factors, such as per capita net income of rural residents, the proportion of the population employed in agriculture, the size and structure of the permanent rural population, local fiscal revenue, and urbanization level, are the main factors that cause changes of rural settlement patterns. The results of this study can serve as a reference for promoting regional rural sustainable development policies and advancing rural spatial governance and comprehensive revitalization.

Modelling Industrial Growth in The Coastal Area of Semarang Metropolitan Region Using GeOBIA

Central Java Province has the manufacturing industry sector as the largest economic contributor. This manufacturing industry sector growth is mainly supported by the strategic location and infrastructures on the northern coast of Central Java Province (Kendal-Semarang-Demak). However, the high increase of industrial built-up area in the coastal area of Semarang Metropolitan Region has an impact on the conversion of coastal land if the growth cannot be controlled through the spatial plan implementation. This study identified the direction and trends of industrial growth in the coastal area of Semarang Metropolitan Region in 2015-2020 and the industrial growth prediction in 2030 for assessing the industrial zones in spatial plan 2011-2031. This research consists the estimation of industrial built-up area growth using the combination of GeOBIA method and manual interpretation of high-resolution satellite imagery (Sentinel-2A) from 2015 to 2020 and assessment of industrial zones in spatial plan based on the industrial growth prediction model in 2030 using the Cellular Automata Markov (CA-Markov). Industrial built-up area tends to grow towards the western and eastern part of the coastal area of Semarang Metropolitan Region based on map distribution and area growth. The results showed more than 20 % of industrial land is not suitable with the spatial plan in 2020 and it predicted will be increased in 2030, it is necessary to increase the efforts to control the spatial use.

Geomorphology and InSAR-Tracked Surface Displacements in an Ice-Rich Yedoma Landscape

Ice-ridge Yedoma terrain is susceptible to vertical surface displacements by thaw and refreeze of ground ice, and geomorphological processes of mass wasting, erosion and sedimentation. Here we explore the relation between a 3 year data set of InSAR measurements of vertical surface displacements during the thaw season, and geomorphological features in an area in the Indigirka Lowlands, Northeast Siberia. The geomorphology is presented in a geomorphological map, based on interpretation of high resolution visible spectrum satellite imagery, field surveys and available data from paleo-environmental research. The main landforms comprise overlapping drained thaw lake basins and lakes, erosion remnants of Late Pleistocene Yedoma deposits, and a floodplain of a high-sinuosity anastomosing river with ancient river terrace remnants. The spatial distribution of drained thaw lake basins and Yedoma erosion remnants in the study area and its surroundings is influenced by neotectonic movements. The 3 years of InSAR measurement include 2 years of high snowfall and extreme river flooding (2017–2018) and 1 year of modest snowfall, early spring and warm summer (2019). The magnitude of surface displacements varies among the years, and show considerable spatial variation. Distinct spatial clusters of displacement trajectories can be discerned, which relate to geomorphological processes and ground ice conditions. Strong subsidence occurred in particular in 2019. In the wet year of 2017, marked heave occurred at Yedoma plateau surfaces, likely by ice accumulation at the top of the permafrost driven by excess precipitation. The spatial variability of surface displacements is high. This is explored by statistical analysis, and is attributed to the interaction of various processes. Next to ground ice volume change, also sedimentation (peat, colluvial deposition) and shrinkage or swelling of soils with changing water content may have contributed. Tussock tundra areas covered by the extreme 2017 and 2018 spring floods show high subsidence rates and an increase of midsummer thaw depths. We hypothesize that increased flood heights along Siberian lowland rivers potentially induce deeper thaw and subsidence on floodplain margins, and also lowers the drainage thresholds of thaw lakes. Both mechanisms tend to increase floodplain area. This may increase CH4 emission from floodplains, but also may enhance carbon storage in floodplain sedimentary environments.

Landslides triggered by the 2018 Lombok earthquake sequence, Indonesia

During July and August 2018, a series of four Mw > 6.0 earthquakes (Event I − 2018.07.28 Mw 6.4, Event II − 2018.08.05 Mw 6.9, Event III − 2018.08.19 Mw 6.3, and Event IV − 2018.08.19 Mw 6.9) struck Lombok Island, Indonesia. A detailed interpretation of satellite imagery reveals that the Lombok earthquake sequence triggered 10,521 landslides with a cumulative extent of 28.0 km2 over an affected area reaching approximately 1850 km2, which are divided into seismic earthflows, topography-based landslides and gully-based landslides. A total of 5621 and 2980 landslides were triggered by Event II and Events III-IV, respectively, and 1920 landslides were concentrated in the 2018.08.10 cloud-covered area. Event II caused most of the landslides. The landslide areas of all the events follow an exponential distribution. The cumulative landslide number (N) and landslide area (A) exhibit a linear logarithmic relationship of Log(N) = −1.341*Log(A) + 3.9. The western section of Mount (Mt.) Rinjani is clearly a landslide-prone area that contains mainly seismic earthflows. Rainfall is a dominant factor influencing the landslide distribution, and generally, higher rainfall corresponds to a higher landslide concentration. Seismic shaking is not a key factor in deposit failure, while in the lower gently sloping section (elevation < 1500 m), the peak ground acceleration (PGA) of landslides increases with a decrease in elevation. Moreover, the top section of the mountain with a steeper slope (Subzones II and V) caused a stronger seismic amplification effect, resulting in numerous landslides.

Alongshore Classification and Morphometric Analysis of Developed Coastal Belts: An Example from Uruguay, South America

Finkl, C.W. and Makowski, C., 2021. Alongshore classification and morphometric analysis of developed coastal belts: An example from Uruguay, South America. Journal of Coastal Research, 37(5), 906–922. Coconut Creek (Florida), ISSN 0749-0208. Approximately half of the 666-km long Uruguayan coastal belt length is characterized by continuums of developed urban centers that are associated with ports and harbors, residential, recreational, military, shore protection, and/or commercial facilities that occur in association with agricultural pursuits. Modification of the Biophysical Cross-shore Classification System (BCCS) to include such features of anthropogenic change along with naturally occurring coastalscapes permits classification of developed and undeveloped coastal belts based on interpretation of satellite imagery. Characterization of alongshore domains was based on conjoining archetypes into catenary sequences that typify coastal belts both alongshore and cross-shore. Based on interpretation of satellite imagery acquired from Google Earth Pro, Uruguayan coastal belts were divided into six separate departments (Rocha, Maldonado, Canelones, Montevideo, San Jose, Colonia) and characterized by the alongshore widths of archetypes that extended from offshore to several kilometers inland. Compilation of coastal belt catenas showed that promontories and headlands, which comprise resistant igneous and metamorphic rocks, function as anchor points along the shore and are widely interspersed by sedimentary (littoral) domains that contain mainland and barrier beaches that are backed by beach-dune couplets, wetlands, lagoonal flats, and developed upland archetypes. Typical alongshore super domains, which were concised from cross-shore archetypical consequences, include the following predominant types of sequences: Barrier-Beach-Dune, Beach-Dune-Lagoon, Beach-Dune-Wetland, Beach-Dune-Upland, Beach-Cliff-Upland, Beach-Wetland-Flat-Lagoon, and Rock-Cliff-Upland. Certain archetypes, such as rock and developed, were prevalent throughout the coastal belt codifications with their representative symbolizations appearing in the variable code sequences of domains and super domains. Morphometric analysis tables were also compiled with alongshore lengths of domains and super domains, as well as percentages of individual department and Uruguayan coastal belts. This examination of the Uruguayan coast showed for the first time that the BCCS can characterize anthropogenically developed units with naturally occurring biophysical features. Interpreted cross-shore transects, alongshore domains, and all-encompassing super domains allow for a comprehensive classification and morphometric analysis of offshore, inshore, and onshore components along Uruguay's natural and developed coastal belts.

GEOLOGIA DE SUPERFÍCIE E ASPECTOS ESTRUTURAIS DA REGIÃO DE RIO TINTO, FAIXA COSTEIRA DA BACIA DA PARAÍBA

Este trabalho apresenta o mapeamento geológico na escala 1:50.000, da região de Rio Tinto, que está localizada na faixa costeira da Bacia da Paraíba, Sub-bacia Miriri. O estudo foi realizado com o apoio de interpretação de imagens de satélite, perfis litológicos de poços, análise de lâminas petrográficas e de padrões geomorfológicos. O embasamento aflora em aproximadamente 10% da área de estudo e é representado por granitóides neoproterozóicos. Cerca de 90% da área é coberta por depósitos Miocênicos-Pleistocênicos e Quaternários das unidades Barreiras, Pós-Barreiras e depósitos litorâneos, de Mangue, terraços Pleistocênicos e sedimentos Elúvio-Coluviais. A análise dos dados cartográficos e de imagens de radar mostrou que a região possui um relevo dominado por tabuleiros e vales formados pela reativação Quaternária do arcabouço estrutural da faixa costeira da bacia. O Rio Mamanguape está inserido na falha homônima com direção ENE-WSW, e a drenagem secundária está encaixada em falhas que possuem direções ENE-WSW, NW-SE e N-S. Os dados dos perfis de poços evidenciaram a ocorrência de grabens e horsts formados pela estruturação tardia da faixa costeira da bacia.

Comparison of Aerosol Optical Depth from MODIS Product Collection 6.1 and AERONET in the Western United States

Recent observations reveal that dust storms are increasing in the western USA, posing imminent risks to public health, safety, and the economy. Much of the observational evidence has been obtained from ground-based platforms and the visual interpretation of satellite imagery from limited regions. Comprehensive satellite-based observations of long-term aerosol records are still lacking. In an effort to develop such a satellite aerosol dataset, we compared and evaluated the Aerosol Optical Depth (AOD) from Deep Blue (DB) and Dark Target (DT) product collection 6.1 with the Aerosol Robotic Network (AERONET) program in the western USA. We examined the seasonal and monthly average number of Moderate Resolution Imaging Spectroradiometer (MODIS) Aqua DB AOD retrievals per 0.1∘ × 0.1∘ from January 2003 to December 2017 across the region’s different topographic, climatic, and land cover conditions. The number of retrievals in the southwest United States was on average greater than 37 days per 90 days for all seasons except summer. Springtime saw the highest number of AOD retrievals across the southwest, consistent with the peak season for synoptic-scale dust events. The majority of Arizona, New Mexico, and western Texas showed the lowest number of retrievals during the monsoon season. The majority of collocating domains of AOD from the Aqua sensor showed a better correlation with AERONET AOD than AOD from Terra, and the correlation coefficients exhibited large regional variability across the study area. The correlation coefficient between the couplings Aqua DB AOD-AERONET AOD and Terra DB AOD-AERONET AOD ranges from 0.1 to 0.94 and 0.001 to 0.94, respectively. In the majority of the sites that exhibited less than a 0.6 correlation coefficient and few matched data points at the nearest single pixel, the correlations gradually improved when the spatial domain increased to a 50 km × 50 km box averaging domain. In general, the majority of the stations revealed significant correlation between MODIS and AERONET AOD at all spatiotemporal aggregating domains, although MODIS generally overestimated AOD compared to AERONET. However, the correlation coefficient in the southwest United States was the lowest and in stations from a higher latitude was the highest. The difference in the brightness of the land surface and the latitudinal differences in the aerosol inputs from the forest fires and solar zenith angles are some of the factors that manifested the latitudinal correlation differences.

Monitoring sea surface oil slicks; a combined approach using advances in satellite radar, AIS and Metocean information

Abstract In recent decades, industrial activity in the marine environment has dramatically increased. Dense offshore infrastructure combined with an increase in shipping activity creates a complex blend of potential oil pollution and environmental risk. Over the last year, CGG Satellite Mapping, supported by the European Space Agency's business applications, Oil Spill Response Limited and key oil and gas companies, created and managed a SeaScope monitoring demonstration service. SeaScope confirms the presence or absence of slicks around infrastructure; reports on the integrity of operations based on slicks observed; identifies the potential origins of pollution slicks or permitted discharge; and has the ability to distinguish between pollution and natural seepage by using routine satellite monitoring. This paper describes how advances in Earth observation radar combined with AIS (Automatic Identification System), metocean data and experts trained in the interpretation of offshore satellite imagery, can be used to regularly monitor sea surface oil slicks and notify offshore operators of anomalous events within the vicinity of offshore infrastructure and high density shipping areas.

ANALISIS PENGELOLAAN HUTAN OLEH MASYARAKAT DAYAK SIANG DI DESA MUWUN KECAMATAN TANAH SIANG KABUPATEN MURUNG RAYA KALIMATAN TENGAH

Murung Raya District in the middle of the island of Borneo makes this widest district in Central Kalimantan quite isolated. Population centers are separated by hills and mountains. Its natural resources are abundant and have not all been utilized, including coal and gold. This study aims to identify the ecological potential, analyze the management system and determine the performance of forest management by Dayak Siang Communities in Muwun Village, Tanah Siang Subdistrict, Murung Raya District, Central Kalimantan Province. The method used was the interpretation of satellite imagery, sampling techniques, interviews, observation, documentation and performance analysis. The performance value of Forest Areas in Muwun Village, Tanah Siang Subdistrict, Murung Raya District, Central Kalimantan, if viewed in terms of area was not good because 57 ha of natural forest has a level of performance worth 172 with less good category, then in rubber plantations and mixed plantations with an area of 217 ha with a performance value of 315 was very good with a very good category. It can be said that forest management in the Forest area in Muwun Village, Tanah Siang Subdistrict, Murung Raya District Central Kalimantan was less successful because it appeared that around 80% of the forest was ecologically very high on rubber farms and from NTFPs, natural forests have less performance value because the level of community’s awareness to maintain and manage natural forests was still minimal. This happened because they have obtained permission from the Murung Raya Regional Government for forest management on the reason that the local community considered the forest as a source of livelihood to survive.

Color models used for the visual interpretation of satellite data

This study presents the possibilities for using different colour models for visual interpretation of satellite imagery. Using the RGB model to visualise different spectral bands as a false colour composite image make it possible for different types of objects and features on the Earth surface to be highlighted and easily discerned based on their specific colour. Examples are shown based on satellite imagery from several free sources, e.g. the USGS’s Earth Explorer, the ESA’s Copernicus Open Access Hub, etc.

UTILIZATION OF LAND IN THE RIVER BORDER AREA IN GROGOL SUB-DISTRICT SUKOHARJO DISTRICT IN 2018

Utilization of land in the river border area in Grogol Sub-District, Sukoharjo District in 2018. The purpose of this research is to : (1) Knowing the Land Use of River Border Areas in Grogol sub-district; (2) Knowing the types and characteristics of buildings that dominate in the River Border Area, Grogol Sub-District in 2018; (3) Knowing the Suitability of River Border Area Land Use in Grogol Sub-District in 2018 with River Border Regulations. This research uses google earth remote sensing image interpretation method with Quickbird satellite imagery, which describes and knows land use, types and characteristics of buildings and the suitability of land use with boundary regulations. The data collection technique used is through interpretation of satellite imagery, direct field observation. The data analysis technique used was data reduction, data presentation, drawing conclusions / verification. The results of this study are : (1) Utilization of land in the river border area in Grogol Sub-District, Sukoharjo District in 2018, that is Bengawan Solo and Kali Samin rivers consisting of 5 land uses, that is Rice Fields, Garden, House, Empty land, and Factory. (2) The type of land use for house buildings with the characteristics of permanent buildings is more dominant. (3) Most of the land use is in accordance with the Regulation of the Minister of Public Works and Housing Number 28 / PRT / M / 2015, because the use of building land in the river border area in Grogol Sub-District is only about 25 percent or about 42 hectares of the 168 hectares of the river border area in Grogol Sub-District in 2018.

Algorithmic justice and groundtruthing the remote mapping of informal settlements: The example of Ho Chi Minh City’s periphery

The significant advances made in interpreting satellite imagery to monitor urban expansion and informal settlements has made important contributions to urban studies and planning. This paper focuses on the under-examined dimensions of how improvements to classifications of urban areas are not only a technical challenge but lie at the society/technology nexus. We examine why three different research groups produced different urban land use classifications of Ho Chi Minh City, Vietnam from remote sensing images. We trace how a confluence of factors including how the technology intersects with field conditions, researcher assumptions and discretionary choices, and institutional norms and agendas shaped the differences in their results. The different spatial facts they produced raises the issue of adapting algorithms for not only technical accuracy but appropriate social use. In the case of detecting informal settlements, our study finds that groundtruthing through fieldwork or collaborative partnerships is needed to not systematically overlook vulnerable populations and misinform urban planning decisions.

Detection of Forest Cover Disturbances Using Synthetic Aperture Radar (SAR) Time Series Data in the Mount Pulag National Park: A Key Biodiversity Area in Luzon, Philippines

Forest cover disturbances continue to occur at a very alarming rate, which greatly contributes to the loss of biodiversity resources. Therefore, an approach to detect and measure historical deforestation would be of utmost importance – particularly in the near real-time forest monitoring and management, biodiversity conservation, and estimation of carbon emissions. This study introduced a simple approach using synthetic aperture radar (SAR) time-series data, primarily the C-band Sentinel-1A (S1A) VH-polarization with a 10-m resolution, in detecting forest cover disturbances from 2016–2019 in Mount Pulag National Park (MPNP), a protected area (PA) within the island of Luzon and a priority biodiversity conservation area. The process involved using codes in Google Earth Engine (GEE) to produce a forest cover disturbance map; a grid-based geospatial and statistical analysis using the geographic information system; and data validation, and determination of possible causes of disturbances using Google Earth Pro (GEP). The results derived from Sentinel-1 data were also compared to existing optical data, which is a Landsat-based generated disturbance map. The output of this paper revealed that a total of 577 forest disturbances, with an equivalent area of 154.58 ha, were detected from a three-year period. For Landsat-generated data, 402 disturbances were detected with an area of 176.07 ha. However, based on the accuracy assessment using positive predictive value (PPV), the radar data produced higher accuracy (80%) compared to the optical data (73%). The visual observation and interpretation of satellite imageries made it clear that the primary drivers of forest cover disturbances in Mount Pulag are agricultural expansion (48%), forest clearings (26%), landslides (12%), slash-and-burn or “kaingin” (12%), and forest fires (2%).

Alongshore Classification and Morphometric Analysis of Coastal Belts: The State of Oregon, USA

Finkl, C.W. and Makowski, C., 2021. Alongshore classification and morphometric analysis of coastal belts: The state of Oregon, USA. Journal of Coastal Research, 37(2), 238–271. Coconut Creek (Florida), ISSN 0749-0208.Classification of the state of Oregon coastal belt (western United States) was based on the interpretation of satellite imagery using the Biophysical Cross-shore Classification System (BCCS). This seven-county, 480-km stretch was subdivided based on coastal archetypes that formed distinct bio-geomorphological and ecological sequences on both alongshore and cross-shore axes. Shore-parallel and shore-perpendicular units were concomitantly determined by cognitive inspection of satellite images, in which cross-shore archetypical sequences were observed in the first instance to have alongshore spread that could be secondarily codified in terms of alongshore catenas. Because shore-normal transects encompass marine, coastal, and terrestrial environments that have a shore-parallel width, it was possible to extrapolate cross-shore classificatory units into alongshore stretches called domains. Characterization of alongshore domains was based on conjoining archetypes into catenary sequences that typify coastal belts both alongshore and cross-shore from adjacent offshore areas to several kilometers inland. Compilation of coastal belt catenas showed that promontories and headlands, which are composed of resistant igneous and metamorphic rocks, function as anchor points along the shore and are interspersed by sedimentary (littoral) coastal belts that contain mainland and barrier beaches backed by dune, wetland, flat (estuaries), or upland archetypes. A consizement of several alongshore domain catenary sequences within a particular coastal segment thus formed an overarching super domain. Typical alongshore super domains consisted of the following sequences: Barrier-Beach-Dune (Ba-Be-Du), Barrier-Beach-Dune-Wetland (Ba-Be-Du-W), Beach-Dune-Upland (Be-Du-U), Beach-Dune-Wetland (Be-Du-W), Beach-Cliff-Upland (Be-Cl-U), Beach-Wetland-Flat (Be-W-F), and Rock-Cliff-Upland (R-Cl-U). Morphometric analysis of each county allowed the determination of specific characteristics for each super domain (e.g., alongshore length, percentage of the county coastal belt, and percentage of the Oregon coastal belt). Furthermore, archetype morphometrics were calculated to show which cross-shore archetypical features were most prevalent along the state's coastal belt. This examination of the Oregon coast showed that the alongshore classification of coastal belts on a regional scale, which is based on cross-shore biophysical interpretations, offers the opportunity to characterize offshore, inshore, and onshore eco-geomorphological features through the use of super domains and morphometric analysis. This approach to coastal classification thus serves as a blueprint for the comprehensive characterization of coastal belts worldwide.

Comparison of Simple Averaging and Latent Class Modeling to Estimate the Area of Land Cover in the Presence of Reference Data Variability

Estimates of the area or percent area of the land cover classes within a study region are often based on the reference land cover class labels assigned by analysts interpreting satellite imagery and other ancillary spatial data. Different analysts interpreting the same spatial unit will not always agree on the land cover class label that should be assigned. Two approaches for accommodating interpreter variability when estimating the area are simple averaging (SA) and latent class modeling (LCM). This study compares agreement between area estimates obtained from SA and LCM using reference data obtained by seven trained, professional interpreters who independently interpreted an annual time series of land cover reference class labels for 300 sampled Landsat pixels. We also compare the variability of the LCM and SA area estimates over different numbers of interpreters and different subsets of interpreters within each interpreter group size, and examine area estimates of three land cover classes (forest, developed, and wetland) and three change types (forest gain, forest loss, and developed gain). Differences between the area estimates obtained from SA and LCM are most pronounced for the estimates of wetland and the three change types. The percent area estimates of these rare classes were usually greater for LCM compared to SA, with the differences between LCM and SA increasing as the number of interpreters providing the reference data increased. The LCM area estimates generally had larger standard deviations and greater ranges over different subsets of interpreters, indicating greater sensitivity to the selection of the individual interpreters who carried out the reference class labeling.

A Brief Review of Graph Convolutional Neural Network Based Learning for classifying remote sensing images

Researchers have recently begun investigating graph Convolutional neural networks for analyzing remotely sensed images. Modelling an image as a graph is a great way to understand its semantic, and applying the convolution on the graph structure have shown promising results in many real-world application. The hope is that such models help to improve the quality of land cover maps produced by the interpretation of satellite imagery. In this paper, we review recently proposed models based on graph convolution neural network architectures and discuss the challenge that could be taken up to improve the results of this technique and how can we use it for transfer learning between model.