Resolution Land Cover Research Articles

The importance of spatial configuration of neighbouring land cover for explanation of surface temperature of individual patches in urban landscapes

ContextRelationships between spatial configuration of urban form and land surface temperature (LST) in the excess heat mitigation context are studied over larger tracts of land not allowing for micro-scale recommendations to urban design.ObjectivesTo identify spatial configuration descriptors (SCDs) of urban form and the size of zone of influence conducive to the formation of the coldest and hottest land cover (LC) patches of different types (buildings, grass, paved and trees) from 2 m resolution LC and 2 and 100 m resolution LST maps at two time-steps in the summer.MethodsRandom Forest regression models were deployed to explain the LST of individual LC patches of different types based on SCDs of core LC patches and patches in their neighbourhoods. ANOVA was used to determine significantly different values of the most important SCDs associated with the coldest and hottest LC patches, and analysis of quartiles informed specification of their ranges.ResultsUrban form in the immediate neighbourhood to the core LC patches had a strong influence on their LST. Low elevation, high proximity to water, and high aggregation of trees, being important to the formation of the coldest patches of all types. High resolution of LST contributed to a higher accuracy of results. Elevation and proximity to water gained in importance as summer progressed.ConclusionsSpatial configuration of urban form in the nearest proximity to individual LC patches and the use of fine resolution LST data are essential for issuing heat mitigation recommendations to urban planners relevant to micro-scales.

Cloudy-sky land surface temperature from VIIRS and MODIS satellite data using a surface energy balance-based method

Land surface temperature (LST) has been effectively retrieved from thermal infrared (TIR) satellite measurements under clear-sky conditions. However, TIR satellite data are often severely contaminated by clouds, which cause spatiotemporal discontinuities and low retrieval accuracy in the LST products. Several solutions have been proposed to fill the “gaps”; however, a majority of these possess constraints. For example, fusion methods with microwave data suffer from coarse spatial resolution and diverse land cover types while spatial-temporal interpolation methods neglect cloudy cooling effects. We developed a novel method to estimate cloudy-sky LST from polar-orbiting satellite data based on the surface energy balance (SEB) principle. First, the hypothetical clear-sky LST of missing or likely cloud-contaminated pixels was reconstructed by assimilating high-quality satellite retrievals into a time-evolving model built from reanalysis data using a Kalman filter data assimilation algorithm. Second, clear-sky LST was hypothetically corrected by accounting for cloud cooling based on SEB theory. The proposed method was applied to Visible Infrared Imaging Radiometer Suite (VIIRS) and Moderate Resolution Imaging Spectroradiometer (MODIS) data, and further validated using ground measurements of fourteen sites from SURFRAD, BSRN, and AmeriFlux in 2013. VIIRS LST recovered from cloud gaps exhibited a root mean square error (RMSE) of 3.54 K, a bias of −0.36 K, R2 of 0.94, and sample size (N) of 2411, comparable to the accuracy of clear-sky LST products and cloudy-sky LST estimation from MODIS (RMSE of 3.69 K, bias of −0.45 K, R2 of 0.93, and N of 2398). Thus, the proposed method performs well across different sensors, seasons, and land cover types. The abnormal retrieval values caused by cloud contamination were also corrected in the proposed method. The overall accuracy was better than the downscaled cloudy-sky LST retrieved from passive microwave (PMW) observations and former SEB-based cloudy-sky LST estimation methods. Validation using time-series measurements showed that the all-sky LST time series, including both clear- and cloudy-sky retrievals, can capture realistic variability without sudden abruptions or discontinuities. RMSE values for the all-sky LST varied from 2.54 to 4.15 K at the fourteen sites. Spatially continuous LST maps over the Contiguous United States were compared with corresponding maps from PMW data in the winter and summer of 2018, exhibiting similar spatial patterns but with additional spatial details. Moreover, sensitivity analysis suggested that the reconstruction of clear-sky LST dominantly impacts the accuracy of cloudy-sky LST estimation. The proposed method can be potentially implemented in similar satellite sensors for global real-time production.

A CALCULATION METHOD OF ECOSYSTEM SERVICE VALUE BASED ON PRODUCT OF NATIONAL GEOGRAPHIC CONDITIONS MONITORING IN CHINA

Abstract. By 2015, Chinese government had completed the project of China's First National Geographic Conditions Census. In this project, high resolution land cover product all over the country had been generated, and would be updated continuously every year. On the basis of this excellent data source, a big data calculating method of land’s ecosystem service value was proposed, in which many other remote sensing information were used too, such as EVI (Enhanced Vegetation Index), NPP (Net Primary Productivity), vegetation growing season data derived from MODIS product. It analyzed the characters of data type, data time phase, and data structure for all the remote sensing information, also the big data’s engendering background and process. A revised ecosystem service value assessment model was used for calculating. Combining the classification system of terrestrial ecosystem in China and the equivalent value factor per unit ecosystem area, the big data calculating algorithm was designed. Shiyan city, Hubei province, China was selected as the study area for validating the calculating method. The results showed that the total ecosystem service value in Shiyan city in 2015 was 1.97 × 1011 CNY, and the per capita ecosystem service value was 5.69 × 104 CNY. Specially, forest supplied the most ecosystem service value which accounted for 78.54 %, followed by water, grassland, farmland, and desert. The research shows that on the basis of multi-source of remote sensing information mainly the high resolution land cover product obtained in the project of China's First National Geographic Conditions Census, high-precision quantification and spatialization ecosystem service value can be calculated and obtained; multi scale spatial display of the calculating results could be achieved to meet different spatial scaling demands; the big data calculating algorithm has solved the problems of design and computation of structured and unstructured big data computing models; the independent research and development software has solved the problem of software requirements, and the operational efficiency and performance can meet the calculating needs.

Global land cover trajectories and transitions

Global land cover (LC) changes threaten sustainability and yet we lack a comprehensive understanding of the gains and losses of LC types, including the magnitudes, locations and timings of transitions. We used a novel, fine-resolution and temporally consistent satellite-derived dataset covering the entire Earth annually from 1992 to 2018 to quantify LC changes across a range of scales. At global and continental scales, the observed trajectories of change for most LC types were fairly smooth and consistent in direction through time. We show these observed trajectories in the context of error margins produced by extrapolating previously published accuracy metrics associated with the LC dataset. For many LC classes the observed changes were found to be within the error margins. However, an important exception was the increase in urban land, which was consistently larger than the error margins, and for which the LC transition was unidirectional. An advantage of analysing the global, fine spatial resolution LC time-series dataset is the ability to identify where and when LC changes have taken place on the Earth. We present LC change maps and trajectories that identify locations with high dynamism, and which pose significant sustainability challenges. We focused on forest loss and urban growth at the national scale, identifying the top 10 countries with the largest percentages of forest loss and urban growth globally. Crucially, we found that most of these ‘worst-case’ countries have stabilized their forest losses, although urban expansion was monotonic in all cases. These findings provide crucial information to support progress towards the UN’s SDGs.

The Imprint of Built-Up Land Expansion on Cropland Distribution and Productivity in Shandong Province

Grain self-sufficiency is a national food security target of China. The way that built-up land expansion impacts upon cropland loss and food provision needs to be explored in the major grain producing areas. Shandong Province is an important agricultural food production region, which is also experiencing rapidly urbanizing. Here we assessed the spatiotemporal distribution of cropland loss due to built-up land expansion and landscape dynamics of cropland during 2000–2020, by using 30 m resolution land cover data. We also analyzed the potential yield change influenced by cropland loss. The results showed that the area of built-up land expanded by 5199 km2 from 2000–2010, and 11,949 km2 from 2010–2020. Approximately 95% of the new built-up land was from cropland during the two stages, and the primary mode of built-up land expansion was the edge expansion. The patch density and the patch size of cropland kept increasing and decreasing, respectively, and the aggregation index kept decreasing from 2000 to 2020, indicating increased cropland fragmentation. The proportion of occupied cropland with potential yield greater than 7500 kg/ha was 25% and 37% during the former and the latter period. Thus, higher quality cropland was encroached in the recent period. The findings could provide meaningful implications for making sustainable land use development strategies in the study area and other similar regions.

Continental-Scale Land Cover Mapping at 10 m Resolution Over Europe (ELC10)

Land cover maps are important tools for quantifying the human footprint on the environment and facilitate reporting and accounting to international agreements addressing the Sustainable Development Goals. Widely used European land cover maps such as CORINE (Coordination of Information on the Environment) are produced at medium spatial resolutions (100 m) and rely on diverse data with complex workflows requiring significant institutional capacity. We present a 10 m resolution land cover map (ELC10) of Europe based on a satellite-driven machine learning workflow that is annually updatable. A random forest classification model was trained on 70K ground-truth points from the LUCAS (Land Use/Cover Area Frame Survey) dataset. Within the Google Earth Engine cloud computing environment, the ELC10 map can be generated from approx. 700 TB of Sentinel imagery within approx. 4 days from a single research user account. The map achieved an overall accuracy of 90% across eight land cover classes and could account for statistical unit land cover proportions within 3.9% (R2 = 0.83) of the actual value. These accuracies are higher than that of CORINE (100 m) and other 10 m land cover maps including S2GLC and FROM-GLC10. Spectro-temporal metrics that capture the phenology of land cover classes were most important in producing high mapping accuracies. We found that the atmospheric correction of Sentinel-2 and the speckle filtering of Sentinel-1 imagery had a minimal effect on enhancing the classification accuracy (<1%). However, combining optical and radar imagery increased accuracy by 3% compared to Sentinel-2 alone and by 10% compared to Sentinel-1 alone. The addition of auxiliary data (terrain, climate and night-time lights) increased accuracy by an additional 2%. By using the centroid pixels from the LUCAS Copernicus module polygons we increased accuracy by <1%, revealing that random forests are robust against contaminated training data. Furthermore, the model requires very little training data to achieve moderate accuracies—the difference between 5K and 50K LUCAS points is only 3% (86% vs. 89%). This implies that significantly less resources are necessary for making in situ survey data (such as LUCAS) suitable for satellite-based land cover classification. At 10 m resolution, the ELC10 map can distinguish detailed landscape features like hedgerows and gardens, and therefore holds potential for aerial statistics at the city borough level and monitoring property-level environmental interventions (e.g., tree planting). Due to the reliance on purely satellite-based input data, the ELC10 map can be continuously updated independent of any country-specific geographic datasets.

Spatial Resolution Enhancement for Large-Scale Land Cover Mapping via Weakly Supervised Deep Learning

Multispectral satellite imagery is the primary data source for monitoring land cover change and characterizing land cover globally. However, the consistency of land cover monitoring is limited by the spatial and temporal resolutions of the acquired satellite images. The public availability of daily high-resolution images is still scarce. This paper aims to fill this gap by proposing a novel spatiotemporal fusion method to enhance daily low spatial resolution land cover mapping using a weakly supervised deep convolutional neural network. We merge Sentinel images and moderate resolution imaging spectroradiometer (MODIS )-derived thematic land cover maps under the application background of massive remote sensing data and the large spatial resolution gaps between MODIS data and Sentinel images. The neural network training was conducted on the public data set SEN12MS, while the validation and testing used ground truth data from the 2020 IEEE Geoscience and Remote Sensing Society data fusion contest. The proposed data fusion method shows that the synthesized land cover map has significantly higher spatial resolution than the corresponding MODIS-derived land cover map. The ensemble approach can be implemented for generating high-resolution time series of satellite images by fusing fine images from Sentinel-1 and -2 and daily coarse images from MODIS.

Shrub decline and expansion of wetland vegetation revealed by very high resolution land cover change detection in the Siberian lowland tundra

Vegetation change, permafrost degradation and their interactions affect greenhouse gas fluxes, hydrology and surface energy balance in Arctic ecosystems. The Arctic shows an overall “greening” trend (i.e. increased plant biomass and productivity) attributed to expansion of shrub vegetation. However, Arctic shrub dynamics show strong spatial variability and locally “browning” may be observed. Mechanistic understanding of greening and browning trends is necessary to accurately assess the response of Arctic vegetation to a changing climate. In this context, the Siberian Arctic is an understudied region. Between 2010 and 2019, increased browning (as derived from the MODIS Enhanced Vegetation Index) was observed in the Eastern Siberian Indigirka Lowlands. To support interpretation of local greening and browning dynamics, we quantified changes in land cover and transition probabilities in a representative tundra site in the Indigirka Lowlands using a timeseries of three very high resolution (VHR) (0.5 m) satellite images acquired between 2010 and 2019. Using spatiotemporal Potts model regularization, we substantially reduced classification errors related to optical and phenological inconsistencies in the image material. VHR images show that recent browning was associated with declines in shrub, lichen and tussock vegetation and increases in open water, sedge and especially Sphagnum vegetation. Observed formation and expansion of small open water bodies in shrub dominated vegetation suggests abrupt thaw of ice-rich permafrost. Transitions from open water to sedge and Sphagnum, indicate aquatic succession upon disturbance. The overall shift towards open water and wetland vegetation suggests a wetting trend, likely associated with permafrost degradation. Landsat data confirmed widespread expansion of surface water throughout the Indigirka Lowlands. However, the increase in the area of small water bodies observed in VHR data was not visible in Landsat-derived surface water data, which suggests that VHR data is essential for early detection of small-scale disturbances and associated vegetation change in permafrost ecosystems.

Deep Subpixel Mapping Based on Semantic Information Modulated Network for Urban Land Use Mapping

Mixed pixel problem is omnipresent in remote sensing images for urban land use interpretation due to the hardware limitations. Subpixel mapping (SPM) is a usual way to solve this problem by improving the observation scale and realizing a finer spatial resolution land cover mapping. Recently, deep learning-based subpixel mapping network (DLSMNet) was proposed, benefited from its strong representation and learning ability, to restore a visually pleasing finer mapping. However, the spatial context features of artifacts are usually aggregated and progressively lost during the forward pass of the network without sufficient representation, which make it difficult to be learned and restored. In this article, a semantic information modulated (SIM) deep subpixel mapping network (SIMNet) is proposed, which uses low-resolution semantic images as prior, to reinforce the representation of spatial context features. In SIMNet, SIM module is proposed to parametrically incorporate the semantic prior into the state-of-the-art (SOTA) feed forward network architecture in an end-to-end training fashion. Furthermore, stacked SIM module with residual blocks (SIM_ResBlock) is adopted to pass the representation of spatial context feature to the deep layers, to get it fully learned during backpropagation. Experiments have been implemented on three public urban scenario data sets, and the SIMNet generates a clearer outline of artificial facilities with sufficient spatial context, and is distinctive for even individual building, which is challenging for other SOTA DLSMNet. The results demonstrate that the proposed SIMNet is a promising way for high-resolution urban land use mapping from easily available lower resolution remote sensing images.Mixed pixel problem is omnipresent in remote sensing images for urban land use interpretation due to the hardware limitations. Subpixel mapping (SPM) is a usual way to solve this problem by improving the observation scale and realizing a finer spatial resolution land cover mapping. Recently, deep learning-based subpixel mapping network (DLSMNet) was proposed, benefited from its strong representation and learning ability, to restore a visually pleasing finer mapping. However, the spatial context features of artifacts are usually aggregated and progressively lost during the forward pass of the network without sufficient representation, which make it difficult to be learned and restored. In this article, a semantic information modulated (SIM) deep subpixel mapping network (SIMNet) is proposed, which uses low-resolution semantic images as prior, to reinforce the representation of spatial context features. In SIMNet, SIM module is proposed to parametrically incorporate the semantic prior into the state-of-the-art (SOTA) feed forward network architecture in an end-to-end training fashion. Furthermore, stacked SIM module with residual blocks (SIM_ResBlock) is adopted to pass the representation of spatial context feature to the deep layers, to get it fully learned during backpropagation. Experiments have been implemented on three public urban scenario data sets, and the SIMNet generates a clearer outline of artificial facilities with sufficient spatial context, and is distinctive for even individual building, which is challenging for other SOTA DLSMNet. The results demonstrate that the proposed SIMNet is a promising way for high-resolution urban land use mapping from easily available lower resolution remote sensing images.

Deep Learning for Feature-Level Data Fusion: Higher Resolution Reconstruction of Historical Landsat Archive

Long-term record of fine spatial resolution remote sensing datasets is critical for monitoring and understanding global environmental change, especially with regard to fine scale processes. However, existing freely available global land surface observations are limited by medium to coarse resolutions (e.g., 30 m Landsat) or short time spans (e.g., five years for 10 m Sentinel-2). Here we developed a feature-level data fusion framework using a generative adversarial network (GAN), a deep learning technique, to leverage the overlapping Landsat and Sentinel-2 observations during 2016–2019, and reconstruct 10 m Sentinel-2 like imagery from 30 m historical Landsat archives. Our tests with both simulated data and actual Landsat/Sentinel-2 imagery showed that the GAN-based fusion method could accurately reconstruct synthetic Landsat data at an effective resolution very close to that of the real Sentinel-2 observations. We applied the GAN-based model to two dynamic systems: (1) land over dynamics including phenology change, cropping rotation, and water inundation; and (2) human landscape changes such as airport construction, coastal expansion, and urbanization, via historical reconstruction of 10 m Landsat observations from 1985 to 2018. The resulting comparison further validated the robustness and efficiency of our proposed framework. Our pilot study demonstrated the promise of transforming 30 m historical Landsat data into a 10 m Sentinel-2-like archive with advanced data fusion. This will enhance Landsat and Sentinel-2 data science, facilitate higher resolution land cover and land use monitoring, and global change research.

Integrating global land cover products to refine GlobeLand30 forest types: a case study of conterminous United States (CONUS)

ABSTRACT Highly accurate and detailed information on land cover products is crucial in studying global climate change and sustainable development. GlobeLand30 is the first global 30 m resolution Land cover (LC) product based on remote sensing data developed by Chinese scientists. GlobeLand30 has 10 first-level classes of products. However, no second-level classification products have been released. This study presents an integration method based on fuzzy theory and combines three 30 m resolution LC products, namely, National Land Cover Database 2011 (NLCD 2011), Fine Resolution Observation and Monitoring of Global Land Cover Segmentation 2010 (FROM-GLC-Seg) and global forest cover data (treecover2010) products, using the European Environment Information and Observation Network Action Group on Land Monitoring in the European (EAGLE) system of semantic translation. The conterminous United States region is adopted as the research area, and the GlobeLand30 (2010) forest class is subdivided into coniferous, broadleaf and mixed forests. Three different weighted voting methods are applied. The difference is whether the user or local accuracy of the source product is considered. The result shows that the respective accuracies of broadleaf, coniferous, and mixed forests are approximately 65.8%, 57.0%, and 40.9% of the weighted voting method without considering any product’s accuracy; 79.3%, 65.9%, and 58.4% of the weighted voting method considering the user accuracy; and 79.9%, 69.9%, and 59.3% of the weighted voting method considering the local accuracy of each product, respectively. The proposed method for refining the GlobeLand30 first class forest can be applied to other classes and land cover products.

Spatiotemporal Variability of Heat Storage in Major U.S. Cities—A Satellite-Based Analysis

Heat storage, ΔQs, is quantified for 10 major U.S. cities using a method called the thermal variability scheme (TVS), which incorporates urban thermal mass parameters and the variability of land surface temperatures. The remotely sensed land surface temperature (LST) is retrieved from the GOES-16 satellite and is used in conjunction with high spatial resolution land cover and imperviousness classes. New York City is first used as a testing ground to compare the satellite-derived heat storage model to two other methods: a surface energy balance (SEB) residual derived from numerical weather model fluxes, and a residual calculated from ground-based eddy covariance flux tower measurements. The satellite determination of ΔQs was found to fall between the residual method predicted by both the numerical weather model and the surface flux stations. The GOES-16 LST was then downscaled to 1-km using the WRF surface temperature output, which resulted in a higher spatial representation of storage heat in cities. The subsequent model was used to predict the total heat stored across 10 major urban areas across the contiguous United States for August 2019. The analysis presents a positive correlation between population density and heat storage, where higher density cities such as New York and Chicago have a higher capacity to store heat when compared to lower density cities such as Houston or Dallas. Application of the TVS ultimately has the potential to improve closure of the urban surface energy balance.

First 1-M Resolution Land Cover Map Labeling the Overlap in the 3rd Dimension: The 2018 Map for Wallonia

Land cover maps contribute to a large diversity of geospatial applications, including but not limited to land management, hydrology, land use planning, climate modeling and biodiversity monitoring. In densely populated and highly fragmented landscapes as observed in the Walloon region (Belgium), very high spatial resolution is required to depict all the infrastructures, buildings and most of the structural elements of the semi-natural landscapes (like hedges and small water bodies). Because of the resolution, the vertical dimension needs explicit handling to avoid discontinuities incompatible with many applications. For example, how to map a river flowing under a bridge? The particularity of our data is to provide a two-digit land cover code to label all the overlapping items. The identification of all the overlaps resulted from the combination of remote sensing image analysis and decision rules involving ancillary data. The final product is therefore semantically precise and accurate in terms of land cover description thanks to the addition of 24 classes on top of the 11 pure land cover classes. The quality of the map has been assessed using a state-of-the-art validation scheme. Its overall accuracy is as high as 91.5%, with an average producer’s accuracy of 86% and an average user’s accuracy of 91%.

Effect of Independent Variables on Urban Flood Models

The simulation accuracy of urban flood models is affected by independent variables describing terrain resolution and artificial land cover. An evaluation of these effects could provide suggestions for the improvement of simulation accuracy when the available terrain resolutions and representation methods of land cover are different. This paper focused on exploring and evaluating these effects on simulation accuracy using two indicators, relative depth accuracy (RDA) and relative area accuracy (RAA). The study area was the Nanjing Jianye district in China, which has experienced extensive urbanization. Designed rainfall (2 and 10 year return periods) and three terrain resolutions (17, 35, and 70 m) were used in this paper. Building blocks (BB), road drainage (RD), and a combination of both (BB + RD) were compared to present the effect of artificial land cover. Real flood events were initially simulated as a model verification case, and hypothetic modeling scenarios were simulated to evaluate the effects of different resolutions and representation methods. The results indicate that the effect of terrain resolutions on simulation accuracy was more obvious than that of artificial land cover in the study area. In this paper, 20–30% higher accuracy could be achieved in the 35 m resolution model with respect to the 70 m resolution model. A relative accuracy of 94% was achieved in the 17 m resolution model when using the BB method, which was 5% higher than that using the RD method. This paper shows that evaluating the effects of terrain resolution and artificial land cover is effective and helpful for improving the simulation accuracy of urban flood models in extensively urbanized districts.

Improving geographical accessibility modeling for operational use by local health actors

BackgroundGeographical accessibility to health facilities remains one of the main barriers to access care in rural areas of the developing world. Although methods and tools exist to model geographic accessibility, the lack of basic geographic information prevents their widespread use at the local level for targeted program implementation. The aim of this study was to develop very precise, context-specific estimates of geographic accessibility to care in a rural district of Madagascar to help with the design and implementation of interventions that improve access for remote populations.MethodsWe used a participatory approach to map all the paths, residential areas, buildings and rice fields on OpenStreetMap (OSM). We estimated shortest routes from every household in the District to the nearest primary health care center (PHC) and community health site (CHS) with the Open Source Routing Machine (OSMR) tool. Then, we used remote sensing methods to obtain a high resolution land cover map, a digital elevation model and rainfall data to model travel speed. Travel speed models were calibrated with field data obtained by GPS tracking in a sample of 168 walking routes. Model results were used to predict travel time to seek care at PHCs and CHSs for all the shortest routes estimated earlier. Finally, we integrated geographical accessibility results into an e-health platform developed with R Shiny.ResultsWe mapped over 100,000 buildings, 23,000 km of footpaths, and 4925 residential areas throughout Ifanadiana district; these data are freely available on OSM. We found that over three quarters of the population lived more than one hour away from a PHC, and 10–15% lived more than 1 h away from a CHS. Moreover, we identified areas in the North and East of the district where the nearest PHC was further than 5 h away, and vulnerable populations across the district with poor geographical access (> 1 h) to both PHCs and CHSs.ConclusionOur study demonstrates how to improve geographical accessibility modeling so that results can be context-specific and operationally actionable by local health actors. The importance of such approaches is paramount for achieving universal health coverage (UHC) in rural areas throughout the world.

US EPA EnviroAtlas Meter-Scale Urban Land Cover (MULC): 1-m Pixel Land Cover Class Definitions and Guidance.

This article defines the land cover classes used in Meter-scale Urban Land Cover (MULC), a unique, high resolution (one meter2 per pixel) land cover dataset developed for 30 US communities for the United States Environmental Protection Agency (US EPA) EnviroAtlas. MULC data categorize the landscape into these land cover classes: impervious surface, tree, grass-herbaceous, shrub, soil-barren, water, wetland and agriculture. MULC data are used to calculate approximately 100 EnviroAtlas metrics that serve as indicators of nature's benefits (ecosystem goods and services). MULC, a dataset for which development is ongoing, is produced by multiple classification methods using aerial photo and LiDAR datasets. The mean overall fuzzy accuracy across the EnviroAtlas communities is 88% and mean Kappa coefficient is 0.84. MULC is available in EnviroAtlas via web browser, web map service (WMS) in the user's geographic information system (GIS), and as downloadable data at EPA Environmental Data Gateway. Fact Sheets and metadata for each MULC Community are available through EnviroAtlas. Some MULC applications include mapping green and grey infrastructure, connecting land cover with socioeconomic/demographic variables, street tree planting, urban heat island analysis, mosquito habitat risk mapping and bikeway planning. This article provides practical guidance for using MULC effectively and developing similar high resolution (HR) land cover data.

Improving 3-m Resolution Land Cover Mapping through Efficient Learning from an Imperfect 10-m Resolution Map

Substantial progress has been made in the field of large-area land cover mapping as the spatial resolution of remotely sensed data increases. However, a significant amount of human power is still required to label images for training and testing purposes, especially in high-resolution (e.g., 3-m) land cover mapping. In this research, we propose a solution that can produce 3-m resolution land cover maps on a national scale without human efforts being involved. First, using the public 10-m resolution land cover maps as an imperfect training dataset, we propose a deep learning based approach that can effectively transfer the existing knowledge. Then, we improve the efficiency of our method through a network pruning process for national-scale land cover mapping. Our proposed method can take the state-of-the-art 10-m resolution land cover maps (with an accuracy of 81.24% for China) as the training data, enable a transferred learning process that can produce 3-m resolution land cover maps, and further improve the overall accuracy (OA) to 86.34% for China. We present detailed results obtained over three mega cities in China, to demonstrate the effectiveness of our proposed approach for 3-m resolution large-area land cover mapping.

Spatio-Temporal Pattern of Land Degradation from 1990 to 2015 in Mongolia

Land degradation is an important environmental problem facing the world. “Land Degradation Neutrality” is one of the core indicators in the 15th goal of the “United Nations Sustainable Development Goals” for 2030. Mongolia is an important country for global land degradation. The increasingly serious land degradation has caused a direct impact on the ecosystem of the entire Mongolian plateau. We analyzed the patterns of land degradation and restoration during 1990–2010 and 2010–2015 and determined the driving forces behind the variations, by using fine resolution land cover data for the first time in Mongolia. The results showed that the spatial distribution of newly increased land degradation and restoration have a strong transitional nature. For the past 25 years, the trend of land change in Mongolia was dominated by land degradation. However, land degradation was accompanied by ongoing restoration of some land areas, and the capacity for land restoration has been gradually improved. This study discovers a series of typical land degradation and restoration regions and provides an interpretation of the driving forces in these areas. The joint effects of natural and socioeconomic factors have been found to result in land degradation and restoration in different regions.

Sub-pixel spectral clustering model of quantum mechanism effect for hyperspectral images

The problem of pixel mixture, which is caused by low spatial resolution and complex land cover, restricts the development of target classification and detection for hyperspectral images (HSIs) in many application fields. In this paper, a novel sub-pixel spectral clustering algorithm of quantum mechanism effect is proposed for hyperspectral remote sensing images. In order to effectively describe the energy distribution of spectral vectors, we define a nonlinear mixture mathematic model with self-consistency for mixed pixel. Considering the self-consistent conditions between lattice models and quantum impurity model, Green function method is used to design the clusters mean field impurities solver, which can decompose the mixed spectral accurately and quickly. Finally, the spectral samples with smaller quantum potential energy and greater local density are applied as cluster center to obtain the final clustering result. The effectiveness of this method is illustrated on several hyperspectral remote sensing images, and experiments show that it can improve the ability of clustering high dimensional non-spherical structure data.

The influence of land cover data on farm-scale valuations of natural capital

The valuation of natural capital within individual farms could inform environmentally beneficial land use change and form the basis of agricultural subsidy schemes based on the provision of ecosystem services. Land cover extents can be used in a benefit transfer approach to produce monetary valuations of natural capital rapidly and at low cost. However, the methodology has not before been used within individual farms, and the impact of land cover data characteristics on the accuracy of valuations is uncertain. Here, we apply the approach to five UK farms of contrasting size, configuration and farming style, using three widely available land cover products. Results show that the land cover product used has a substantial impact on valuations, with differences of up to 58%, and the magnitude of this effect varies considerably according to the landscape structure of the farm. At most sites, valuation differences are driven by the extent of woodland recorded in the landscape, with higher resolution land cover products incorporating larger amounts of woodland through inclusion of smaller patches, leading to higher overall valuations. Integrating more accurate land cover data and accounting for the condition, configuration and location of natural capital has potential to improve the accuracy of valuations.