An effective approach for gap-filling continental scale remotely sensed time-series

Human movement and environmental barriers shape the emergence of dengue

Understanding how emerging infectious diseases spread within and between countries is essential to contain future pandemics. Spread to new areas requires connectivity between one or more sources and a suitable local environment, but how these two factors interact at different stages of disease emergence remains largely unknown. Further, no analytical framework exists to examine their roles. Here we develop a dynamic modelling approach for infectious diseases that explicitly models both connectivity via human movement and environmental suitability interactions. We apply it to better understand recently observed (1995-2019) patterns as well as predict past unobserved (1983-2000) and future (2020-2039) spread of dengue in Mexico and Brazil. We find that these models can accurately reconstruct long-term spread pathways, determine historical origins, and identify specific routes of invasion. We find early dengue invasion is more heavily influenced by environmental factors, resulting in patchy non-contiguous spread, while short and long-distance connectivity becomes more important in later stages. Our results have immediate practical applications for forecasting and containing the spread of dengue and emergence of new serotypes. Given current and future trends in human mobility, climate, and zoonotic spillover, understanding the interplay between connectivity and environmental suitability will be increasingly necessary to contain emerging and re-emerging pathogens.

Understanding transmission risk and predicting environmental suitability for Mayaro Virus in Central and South America

Mayaro virus (MAYV) is a mosquito-borne Alphavirus that is widespread in South America. MAYV infection often presents with non-specific febrile symptoms but may progress to debilitating chronic arthritis or arthralgia. Despite the pandemic threat of MAYV, its true distribution remains unknown. The objective of this study was to clarify the geographic distribution of MAYV using an established risk mapping framework. This consisted of generating evidence consensus scores for MAYV presence, modeling the potential distribution of MAYV in select countries across Central and South America, and estimating the population residing in areas suitable for MAYV transmission. We compiled a georeferenced compendium of MAYV occurrence in humans, animals, and arthropods. Based on an established evidence consensus framework, we integrated multiple information sources to assess the total evidence supporting ongoing transmission of MAYV within each country in our study region. We then developed high resolution maps of the disease's estimated distribution using a boosted regression tree approach. Models were developed using nine climatic and environmental covariates that are related to the MAYV transmission cycle. Using the output of our boosted regression tree models, we estimated the total population living in regions suitable for MAYV transmission. The evidence consensus scores revealed high or very high evidence of MAYV transmission in several countries including Brazil (especially the states of Mato Grosso and Goiás), Venezuela, Peru, Trinidad and Tobago, and French Guiana. According to the boosted regression tree models, a substantial region of South America is suitable for MAYV transmission, including north and central Brazil, French Guiana, and Suriname. Some regions (e.g., Guyana) with only moderate evidence of known transmission were identified as highly suitable for MAYV. We estimate that approximately 58.9 million people (95% CI: 21.4-100.4) in Central and South America live in areas that may be suitable for MAYV transmission, including 46.2 million people (95% CI: 17.6-68.9) in Brazil. Our results may assist in prioritizing high-risk areas for vector control, human disease surveillance and ecological studies.

Updating estimates of Plasmodium knowlesi malaria risk in response to changing land use patterns across Southeast Asia

BACKGROUND

Plasmodium knowlesi is a zoonotic parasite that causes malaria in humans. The pathogen has a natural host reservoir in certain macaque species and is transmitted to humans via mosquitoes of the Anopheles Leucosphyrus Group. The risk of human P. knowlesi infection varies across Southeast Asia and is dependent upon environmental factors. Understanding this geographic variation in risk is important both for enabling appropriate diagnosis and treatment of the disease and for improving the planning and evaluation of malaria elimination. However, the data available on P. knowlesi occurrence are biased towards regions with greater surveillance and sampling effort. Predicting the spatial variation in risk of P. knowlesi malaria requires methods that can both incorporate environmental risk factors and account for spatial bias in detection.

METHODS & RESULTS

We extend and apply an environmental niche modelling framework as implemented by a previous mapping study of P. knowlesi transmission risk which included data up to 2015. We reviewed the literature from October 2015 through to March 2020 and identified 264 new records of P. knowlesi, with a total of 524 occurrences included in the current study following consolidation with the 2015 study. The modelling framework used in the 2015 study was extended, with changes including the addition of new covariates to capture the effect of deforestation and urbanisation on P. knowlesi transmission.

DISCUSSION

Our map of P. knowlesi relative transmission suitability estimates that the risk posed by the pathogen is highest in Malaysia and Indonesia, with localised areas of high risk also predicted in the Greater Mekong Subregion, The Philippines and Northeast India. These results highlight areas of priority for P. knowlesi surveillance and prospective sampling to address the challenge the disease poses to malaria elimination planning.

Updating estimates of Plasmodium knowlesi malaria risk in response to changing land use patterns across Southeast Asia

Background

Plasmodium knowlesi is a zoonotic parasite that causes malaria in humans. The pathogen has a natural host reservoir in certain macaque species and is transmitted to humans via mosquitoes of the Anopheles Leucosphyrus Group. The risk of human P. knowlesi infection varies across Southeast Asia and is dependent upon environmental factors. Understanding this geographic variation in risk is important both for enabling appropriate diagnosis and treatment of the disease and for improving the planning and evaluation of malaria elimination. However, the data available on P. knowlesi occurrence are biased towards regions with greater surveillance and sampling effort. Predicting the spatial variation in risk of P. knowlesi malaria requires methods that can both incorporate environmental risk factors and account for spatial bias in detection.

Methods & Results

We extend and apply an environmental niche modelling framework as implemented by a previous mapping study of P. knowlesi transmission risk which included data up to 2015. We reviewed the literature from October 2015 through to March 2020 and identified 264 new records of P. knowlesi, with a total of 524 occurrences included in the current study following consolidation with the 2015 study. The modelling framework used in the 2015 study was extended, with changes including the addition of new covariates to capture the effect of deforestation and urbanisation on P. knowlesi transmission.

Discussion

Our map of P. knowlesi relative transmission suitability estimates that the risk posed by the pathogen is highest in Malaysia and Indonesia, with localised areas of high risk also predicted in the Greater Mekong Subregion, The Philippines and Northeast India. These results highlight areas of priority for P. knowlesi surveillance and prospective sampling to address the challenge the disease poses to malaria elimination planning.

Machine learning based estimation of field-scale daily, high resolution, multi-depth soil moisture for the Western and Midwestern United States

Background

High-resolution soil moisture estimates are critical for planning water management and assessing environmental quality. In-situ measurements alone are too costly to support the spatial and temporal resolutions needed for water management. Recent efforts have combined calibration data with machine learning algorithms to fill the gap where high resolution moisture estimates are lacking at the field scale. This study aimed to provide calibrated soil moisture models and methodology for generating gridded estimates of soil moisture at multiple depths, according to user-defined temporal periods, spatial resolution and extent.

Methods

We applied nearly one million national library soil moisture records from over 100 sites, spanning the U.S. Midwest and West, to build Quantile Random Forest (QRF) calibration models. The QRF models were built on covariates including soil moisture estimates from North American Land Data Assimilation System (NLDAS), soil properties, climate variables, digital elevation models, and remote sensing-derived indices. We also explored an alternative approach that adopted a regionalized calibration dataset for the Western U.S. The broad-scale QRF models were independently validated according to sampling depths, land cover type, and observation period. We then explored the model performance improved with local samples used for spiking. Finally, the QRF models were applied to estimate soil moisture at the field scale where evaluation was carried out to check estimated temporal and spatial patterns.

Results

The broad-scale QRF model showed moderate performance (R² = 0.53, RMSE = 0.078 m³/m³) when data points from all depth layers (up to 100 cm) were considered for an independent validation. Elevation, NLDAS-derived moisture, soil properties, and sampling depth were ranked as the most important covariates. The best model performance was observed for forest and pasture sites (R² > 0.5; RMSE < 0.09 m³/m³), followed by grassland and cropland (R² > 0.4; RMSE < 0.11 m³/m³). Model performance decreased with sampling depths and was slightly lower during the winter months. Spiking the national QRF model with local samples improved model performance by reducing the RMSE to less than 0.05 m³/m³ for grassland sites. At the field scale, model estimates illustrated more accurate temporal trends for surface than subsurface soil layers. Model estimated spatial patterns need to be further improved and validated with management data.

Conclusions

The model accuracy for top 0-20 cm soil depth (R² > 0.5, RMSE < 0.08 m³/m³) showed promise for adopting the methodology for soil moisture monitoring. The success of spiking the national model with local samples showed the need to collect multi-year high frequency (e.g., hourly) sensor-based field measurements to improve estimates of soil moisture for a longer time period. Future work should improve model performance for deeper depths with additional hydraulic properties and use of locally-selected calibration datasets.

Spatially Explicit River Basin Models for Cost-Benefit Analyses to Optimize Land Use

Recently, a wide range of models have been used in analyzing the costs and benefits of land utilization in river basins. Despite these advances, there is not enough information on how to select appropriate models to perform cost-benefit analyses. A literature search in the Web of Science (WOS) online database was implemented and resulted in the selection of 27 articles that utilized models to perform cost-benefit analyses of river basins. The models reviewed in these papers were categorized into five types: process-based, statistical, probabilistic, data-driven, and modeling frameworks or integrated models. Twenty-six models were reviewed based on their data and input variable needs and user convenience. A SWOT analysis was also performed to highlight the strengths, weaknesses, opportunities, and threats of these models. One of the main strengths is their ability to perform scenario-based analyses while the main drawback is the limited availability of data impeding the use of the models. We found that, to some extent, there is an increase in model applicability as the number of input variables increases but there are exceptions to this observation. Future studies should explicitly report on the necessary time needed for data collection, model development and/or training, and model application. This information is highly valuable to users and modelers when choosing which model to use in performing a particular cost-benefit analysis. These models can be developed and applied to assist sustainable development as well as the sustainable utilization of agricultural parcels within a river basin, which can eventually reduce the negative impacts of intensive agriculture and minimize habitat degradation on water resources.

Multi-Season Phenology Mapping of Nile Delta Croplands Using Time Series of Sentinel-2 and Landsat 8 Green LAI

Space-based cropland phenology monitoring substantially assists agricultural managing practices and plays an important role in crop yield predictions. Multitemporal satellite observations allow analyzing vegetation seasonal dynamics over large areas by using vegetation indices or by deriving biophysical variables. The Nile Delta represents about half of all agricultural lands of Egypt. In this region, intensifying farming systems are predominant and multi-cropping rotations schemes are increasing, requiring a high temporal and spatial resolution monitoring for capturing successive crop growth cycles. This study presents a workflow for cropland phenology characterization and mapping based on time series of green Leaf Area Index (LAI) generated from NASA’s Harmonized Landsat 8 (L8) and Sentinel-2 (S2) surface reflectance dataset from 2016 to 2019. LAI time series were processed for each satellite dataset, which were used separately and combined to identify seasonal dynamics for a selection of crop types (wheat, clover, maize and rice). For the combination of L8 with S2 LAI products, we proposed two time series smoothing and fitting methods: (1) the Savitzky–Golay (SG) filter and (2) the Gaussian Processes Regression (GPR) fitting function. Single-sensor and L8-S2 combined LAI time series were used for the calculation of key crop Land Surface Phenology (LSP) metrics (start of season, end of season, length of season), whereby the detection of cropland growing seasons was based on two established threshold methods, i.e., a seasonal or a relative amplitude value. Overall, the developed phenology extraction scheme enabled identifying up to two successive crop cycles within a year, with a superior performance observed for the seasonal than for the relative threshold method, in terms of consistency and cropland season detection capability. Differences between the time series collections were analyzed by comparing the phenology metrics per crop type and year. Results suggest that L8-S2 combined LAI data streams with GPR led to a more precise detection of the start and end of growing seasons for most crop types, reaching an overall detection of 74% over the total planted crops versus 69% with S2 and 63% with L8 alone. Finally, the phenology mapping allowed us to evaluate the spatial and temporal evolution of the croplands over the agroecosystem in the Nile Delta.

Monitoring Cropland Phenology on Google Earth Engine Using Gaussian Process Regression

Monitoring cropland phenology from optical satellite data remains a challenging task due to the influence of clouds and atmospheric artifacts. Therefore, measures need to be taken to overcome these challenges and gain better knowledge of crop dynamics. The arrival of cloud computing platforms such as Google Earth Engine (GEE) has enabled us to propose a Sentinel-2 (S2) phenology end-to-end processing chain. To achieve this, the following pipeline was implemented: (1) the building of hybrid Gaussian Process Regression (GPR) retrieval models of crop traits optimized with active learning, (2) implementation of these models on GEE (3) generation of spatiotemporally continuous maps and time series of these crop traits with the use of gap-filling through GPR fitting, and finally, (4) calculation of land surface phenology (LSP) metrics such as the start of season (SOS) or end of season (EOS). Overall, from good to high performance was achieved, in particular for the estimation of canopy-level traits such as leaf area index (LAI) and canopy chlorophyll content, with normalized root mean square errors (NRMSE) of 9% and 10%, respectively. By means of the GPR gap-filling time series of S2, entire tiles were reconstructed, and resulting maps were demonstrated over an agricultural area in Castile and Leon, Spain, where crop calendar data were available to assess the validity of LSP metrics derived from crop traits. In addition, phenology derived from the normalized difference vegetation index (NDVI) was used as reference. NDVI not only proved to be a robust indicator for the calculation of LSP metrics, but also served to demonstrate the good phenology quality of the quantitative trait products. Thanks to the GEE framework, the proposed workflow can be realized anywhere in the world and for any time window, thus representing a shift in the satellite data processing paradigm. We anticipate that the produced LSP metrics can provide meaningful insights into crop seasonal patterns in a changing environment that demands adaptive agricultural production.

Determining Temporal Uncertainty of a Global Inland Surface Water Time Series

Earth observation time series are well suited to monitor global surface dynamics. However, data products that are aimed at assessing large-area dynamics with a high temporal resolution often face various error sources (e.g., retrieval errors, sampling errors) in their acquisition chain. Addressing uncertainties in a spatiotemporal consistent manner is challenging, as extensive high-quality validation data is typically scarce. Here we propose a new method that utilizes time series inherent information to assess the temporal interpolation uncertainty of time series datasets. For this, we utilized data from the DLR-DFD Global WaterPack (GWP), which provides daily information on global inland surface water. As the time series is primarily based on optical MODIS (Moderate Resolution Imaging Spectroradiometer) images, the requirement of data gap interpolation due to clouds constitutes the main uncertainty source of the product. With a focus on different temporal and spatial characteristics of surface water dynamics, seven auxiliary layers were derived. Each layer provides probability and reliability estimates regarding water observations at pixel-level. This enables the quantification of uncertainty corresponding to the full spatiotemporal range of the product. Furthermore, the ability of temporal layers to approximate unknown pixel states was evaluated for stratified artificial gaps, which were introduced into the original time series of four climatologic diverse test regions. Results show that uncertainty is quantified accurately (>90%), consequently enhancing the product’s quality with respect to its use for modeling and the geoscientific community.

A Review of Reconstructing Remotely Sensed Land Surface Temperature under Cloudy Conditions

Land surface temperature (LST) is an important environmental parameter in climate change, urban heat islands, drought, public health, and other fields. Thermal infrared (TIR) remote sensing is the main method used to obtain LST information over large spatial scales. However, cloud cover results in many data gaps in remotely sensed LST datasets, greatly limiting their practical applications. Many studies have sought to fill these data gaps and reconstruct cloud-free LST datasets over the last few decades. This paper reviews the progress of LST reconstruction research. A bibliometric analysis is conducted to provide a brief overview of the papers published in this field. The existing reconstruction algorithms can be grouped into five categories: spatial gap-filling methods, temporal gap-filling methods, spatiotemporal gap-filling methods, multi-source fusion-based gap-filling methods, and surface energy balance-based gap-filling methods. The principles, advantages, and limitations of these methods are described and discussed. The applications of these methods are also outlined. In addition, the validation of filled LST values’ cloudy pixels is an important concern in LST reconstruction. The different validation methods applied for reconstructed LST datasets are also reviewed herein. Finally, prospects for future developments in LST reconstruction are provided.

Sentinel-1 Time Series for Crop Identification in the Framework of the Future CAP Monitoring

In this upcoming Common Agricultural Policy (CAP) reform, the use of satellite imagery is taking an increasing role for improving the Integrated Administration and Control System (IACS). Considering the operational aspect of the CAP monitoring process, the use of Sentinel-1 SAR (Synthetic Aperture Radar) images is highly relevant, especially in regions with a frequent cloud cover, such as Belgium. Indeed, SAR imagery does not depend on sunlight and is barely affected by the presence of clouds. Moreover, the SAR signal is particularly sensitive to the geometry and the water content of the target. Crop identification is often a pre-requisite to monitor agriculture at parcel level (ploughing, harvest, grassland mowing, intercropping, etc.) The main goal of this study is to assess the performances and constraints of a SAR-based crop classification in an operational large-scale application. The Random Forest object-oriented classification model is built on Sentinel-1 time series from January to August 2020 only. It can identify crops in the Walloon Region (south part of Belgium) with high performance: 93.4% of well-classified area, representing 88.4% of the parcels. Among the 48 crop groups, the six most represented ones get a F1-score higher or equal to 84%. Additionally, this research documents how the classification performance is affected by different parameters: the SAR orbit, the size of the training dataset, the use of different internal buffers on parcel polygons before signal extraction, the set of explanatory variables, and the period of the time series. In an operational context, this allows to choose the right balance between classification accuracy and model complexity. A key result is that using a training dataset containing only 3.2% of the total number of parcels allows to correctly classify 91.7% of the agricultural area. The impact of rain and snow is also discussed. Finally, this research analyses how the classification accuracy depends on some characteristics of the parcels like their shape or size. This allows to assess the relevance of the classification depending on those characteristics, as well as to identify a subset of parcels for which the global accuracy is higher.

Development of Integrated Crop Drought Index by Combining Rainfall, Land Surface Temperature, Evapotranspiration, Soil Moisture, and Vegetation Index for Agricultural Drought Monitoring

Various drought indices have been used for agricultural drought monitoring, such as Standardized Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI), Palmer Drought Severity Index (PDSI), Soil Water Deficit Index (SWDI), Normalized Difference Vegetation Index (NDVI), Vegetation Health Index (VHI), Vegetation Drought Response Index (VegDRI), and Scaled Drought Condition Index (SDCI). They incorporate such factors as rainfall, land surface temperature (LST), potential evapotranspiration (PET), soil moisture content (SM), and vegetation index to express the meteorological and agricultural aspects of drought. However, these five factors should be combined more comprehensively and reasonably to explain better the dryness/wetness of land surface and the association with crop yield. This study aims to develop the Integrated Crop Drought Index (ICDI) by combining the weather factors (rainfall and LST), hydrological factors (PET and SM), and a vegetation factor (enhanced vegetation index (EVI)) to better express the wet/dry state of land surface and healthy/unhealthy state of vegetation together. The study area was the State of Illinois, a key region of the U.S. Corn Belt, and the quantification and analysis of the droughts were conducted on a county scale for 2004–2019. The performance of the ICDI was evaluated through the comparisons with SDCI and VegDRI, which are the representative drought index in terms of the composite of the dryness and vegetation elements. The ICDI properly expressed both the dry and wet trend of the land surface and described the state of the agricultural drought accompanied by yield damage. The ICDI had higher positive correlations with the corn yields than SDCI and VegDRI during the crucial growth period from June to August for 2004–2019, which means that the ICDI could reflect the agricultural drought well in terms of the dryness/wetness of land surface and the association with crop yield. Future work should examine the other factors for ICDI, such as locality, crop type, and the anthropogenic impacts, on drought. It is expected that the ICDI can be a viable option for agricultural drought monitoring and yield management.

Reconstruction of the Daily MODIS Land Surface Temperature Product Using the Two-Step Improved Similar Pixels Method

The MODIS land surface temperature (LST) product is one of the most widely used data sources to study the climate and energy-water cycle at a global scale. However, the large number of invalid values caused by cloud cover limits the wide application of the MODIS LST. In this study, a two-step improved similar pixels (TISP) method was proposed for cloudy sky LST reconstruction. The TISP method was validated using a temperature-based method over various land cover types. The ground measurements were collected at fifteen stations from 2013 to 2018 during the Heihe Watershed Allied Telemetry Experimental Research (HiWATER) field campaign in China. The estimated theoretical clear-sky temperature indicates that clouds cool the land surface during the daytime and warm it at nighttime. For bare land, the surface temperature shows a clear seasonal trend and very similar across stations, with a cooling amplitude of 4.14 K in the daytime and a warming amplitude of 3.99 K at nighttime, as a yearly average. The validation result showed that the reconstructed LST is highly consistent with in situ measurements and comparable with MODIS LST validation accuracy, with a mean bias of 0.15 K at night (−0.43 K in the day), mean RMSE of 2.91 K at night (4.41 K in the day), and mean R2 of 0.93 at night (0.90 in the day). The developed method maximizes the potential of obtaining quality MODIS LST retrievals, ancillary data, and in situ observations, and the results show high accuracy for most land cover types.

Worldwide continuous gap-filled MODIS land surface temperature dataset

Satellite land surface temperature (LST) is vital for climatological and environmental studies. However, LST datasets are not continuous in time and space mainly due to cloud cover. Here we combine LST with Climate Forecast System Version 2 (CFSv2) modeled temperatures to derive a continuous gap filled global LST dataset at a spatial resolution of 1 km. Temporal Fourier analysis is used to derive the seasonality (climatology) on a pixel-by-pixel basis, for LST and CFSv2 temperatures. Gaps are filled by adding the CFSv2 temperature anomaly to climatological LST. The accuracy is evaluated in nine regions across the globe using cloud-free LST (mean values: R2 = 0.93, Root Mean Square Error (RMSE) = 2.7 °C, Mean Absolute Error (MAE) = 2.1 °C). The provided dataset contains day, night, and daily mean LST for the Eastern Mediterranean. We provide a Google Earth Engine code and a web app that generates gap filled LST in any part of the world, alongside a pixel-based evaluation of the data in terms of MAE, RMSE and Pearson's r.

Green LAI Mapping and Cloud Gap-Filling Using Gaussian Process Regression in Google Earth Engine

For the last decade, Gaussian process regression (GPR) proved to be a competitive machine learning regression algorithm for Earth observation applications, with attractive unique properties such as band relevance ranking and uncertainty estimates. More recently, GPR also proved to be a proficient time series processor to fill up gaps in optical imagery, typically due to cloud cover. This makes GPR perfectly suited for large-scale spatiotemporal processing of satellite imageries into cloud-free products of biophysical variables. With the advent of the Google Earth Engine (GEE) cloud platform, new opportunities emerged to process local-to-planetary scale satellite data using advanced machine learning techniques and convert them into gap-filled vegetation properties products. However, GPR is not yet part of the GEE ecosystem. To circumvent this limitation, this work proposes a general adaptation of GPR formulation to parallel processing framework and its integration into GEE. To demonstrate the functioning and utility of the developed workflow, a GPR model predicting green leaf area index (LAIG) from Sentinel-2 imagery was imported. Although by running this GPR model into GEE any corner of the world can be mapped into LAIG at a resolution of 20 m, here we show some demonstration cases over western Europe with zoom-ins over Spain. Thanks to the computational power of GEE, the mapping takes place on-the-fly. Additionally, a GPR-based gap filling strategy based on pre-optimized kernel hyperparameters is also put forward for the generation of multi-orbit cloud-free LAIG maps with an unprecedented level of detail, and the extraction of regularly-sampled LAIG time series at a pixel level. The ability to plugin a locally-trained GPR model into the GEE framework and its instant processing opens up a new paradigm of remote sensing image processing.

A Machine Learning Approach for Remote Sensing Data Gap-Filling with Open-Source Implementation: An Example Regarding Land Surface Temperature, Surface Albedo and NDVI

Satellite remote sensing has now become a unique tool for continuous and predictable monitoring of geosystems at various scales, observing the dynamics of different geophysical parameters of the environment. One of the essential problems with most satellite environmental monitoring methods is their sensitivity to atmospheric conditions, in particular cloud cover, which leads to the loss of a significant part of data, especially at high latitudes, potentially reducing the quality of observation time series until it is useless. In this paper, we present a toolbox for filling gaps in remote sensing time-series data based on machine learning algorithms and spatio-temporal statistics. The first implemented procedure allows us to fill gaps based on spatial relationships between pixels, obtained from historical time-series. Then, the second procedure is dedicated to filling the remaining gaps based on the temporal dynamics of each pixel value. The algorithm was tested and verified on Sentinel-3 SLSTR and Terra MODIS land surface temperature data and under different geographical and seasonal conditions. As a result of validation, it was found that in most cases the error did not exceed 1 °C. The algorithm was also verified for gaps restoration in Terra MODIS derived normalized difference vegetation index and land surface broadband albedo datasets. The software implementation is Python-based and distributed under conditions of GNU GPL 3 license via public repository.

Gap-Free Monitoring of Annual Mangrove Forest Dynamics in Ca Mau Province, Vietnamese Mekong Delta, Using the Landsat-7-8 Archives and Post-Classification Temporal Optimization

Ecosystem services offered by mangrove forests are facing severe risks, particularly through land use change driven by human development. Remote sensing has become a primary instrument to monitor the land use dynamics surrounding mangrove ecosystems. Where studies formerly relied on bi-temporal assessments of change, the practical limitations concerning data-availability and processing power are slowly disappearing with the onset of high-performance computing (HPC) and cloud-computing services, such as in the Google Earth Engine (GEE). This paper combines the capabilities of GEE, including its entire Landsat-7 and Landsat-8 archives and state-of-the-art classification approaches, with a post-classification temporal analysis to optimize land use classification results into gap-free and consistent information. The results demonstrate its application and value to uncover the spatio-temporal dynamics of mangrove forests and land use changes in Ngoc Hien District, Ca Mau province, Vietnamese Mekong delta. The combination of repeated GEE classification output and post-classification optimization provides valid spatial classification (94–96% accuracy) and temporal interpolation (87–92% accuracy). The findings reveal that the net change of mangroves forests over the 2001–2019 period equals −0.01% annually. The annual gap-free maps enable spatial identification of hotspots of mangrove forest changes, including deforestation and degradation. Post-classification temporal optimization allows for an exploitation of temporal patterns to synthesize and enhance independent classifications towards more robust gap-free spatial maps that are temporally consistent with logical land use transitions. The study contributes to a growing body of work advocating full exploitation of temporal information in optimizing land cover classification and demonstrates its use for mangrove forest monitoring.

Combination of Landsat 8 OLI and Sentinel-1 SAR Time-Series Data for Mapping Paddy Fields in Parts of West and Central Java Provinces, Indonesia

The rise of Google Earth Engine, a cloud computing platform for spatial data, has unlocked seamless integration for multi-sensor and multi-temporal analysis, which is useful for the identification of land-cover classes based on their temporal characteristics. Our study aims to employ temporal patterns from monthly-median Sentinel-1 (S1) C-band synthetic aperture radar data and cloud-filled monthly spectral indices, i.e., Normalized Difference Vegetation Index (NDVI), Modified Normalized Difference Water Index (MNDWI), and Normalized Difference Built-up Index (NDBI), from Landsat 8 (L8) OLI for mapping rice cropland areas in the northern part of Central Java Province, Indonesia. The harmonic function was used to fill the cloud and cloud-masked values in the spectral indices from Landsat 8 data, and smile Random Forests (RF) and Classification And Regression Trees (CART) algorithms were used to map rice cropland areas using a combination of monthly S1 and monthly harmonic L8 spectral indices. An additional terrain variable, Terrain Roughness Index (TRI) from the SRTM dataset, was also included in the analysis. Our results demonstrated that RF models with 50 (RF50) and 80 (RF80) trees yielded better accuracy for mapping the extent of paddy fields, with user accuracies of 85.65% (RF50) and 85.75% (RF80), and producer accuracies of 91.63% (RF80) and 93.48% (RF50) (overall accuracies of 92.10% (RF80) and 92.47% (RF50)), respectively, while CART yielded a user accuracy of only 84.83% and a producer accuracy of 80.86%. The model variable importance in both RF50 and RF80 models showed that vertical transmit and horizontal receive (VH) polarization and harmonic-fitted NDVI were identified as the top five important variables, and the variables representing February, April, June, and December contributed more to the RF model. The detection of VH and NDVI as the top variables which contributed up to 51% of the Random Forest model indicated the importance of the multi-sensor combination for the identification of paddy fields.

Global estimation of anti-malarial drug effectiveness for the treatment of uncomplicated Plasmodium falciparum malaria 1991-2019

BACKGROUND

Anti-malarial drugs play a critical role in reducing malaria morbidity and mortality, but their role is mediated by their effectiveness. Effectiveness is defined as the probability that an anti-malarial drug will successfully treat an individual infected with malaria parasites under routine health care delivery system. Anti-malarial drug effectiveness (AmE) is influenced by drug resistance, drug quality, health system quality, and patient adherence to drug use; its influence on malaria burden varies through space and time.

METHODS

This study uses data from 232 efficacy trials comprised of 86,776 infected individuals to estimate the artemisinin-based and non-artemisinin-based AmE for treating falciparum malaria between 1991 and 2019. Bayesian spatiotemporal models were fitted and used to predict effectiveness at the pixel-level (5 km × 5 km). The median and interquartile ranges (IQR) of AmE are presented for all malaria-endemic countries.

RESULTS

The global effectiveness of artemisinin-based drugs was 67.4% (IQR: 33.3-75.8), 70.1% (43.6-76.0) and 71.8% (46.9-76.4) for the 1991-2000, 2006-2010, and 2016-2019 periods, respectively. Countries in central Africa, a few in South America, and in the Asian region faced the challenge of lower effectiveness of artemisinin-based anti-malarials. However, improvements were seen after 2016, leaving only a few hotspots in Southeast Asia where resistance to artemisinin and partner drugs is currently problematic and in the central Africa where socio-demographic challenges limit effectiveness. The use of artemisinin-based combination therapy (ACT) with a competent partner drug and having multiple ACT as first-line treatment choice sustained high levels of effectiveness. High levels of access to healthcare, human resource capacity, education, and proximity to cities were associated with increased effectiveness. Effectiveness of non-artemisinin-based drugs was much lower than that of artemisinin-based with no improvement over time: 52.3% (17.9-74.9) for 1991-2000 and 55.5% (27.1-73.4) for 2011-2015. Overall, AmE for artemisinin-based and non-artemisinin-based drugs were, respectively, 29.6 and 36% below clinical efficacy as measured in anti-malarial drug trials.

CONCLUSIONS

This study provides evidence that health system performance, drug quality and patient adherence influence the effectiveness of anti-malarials used in treating uncomplicated falciparum malaria. These results provide guidance to countries' treatment practises and are critical inputs for malaria prevalence and incidence models used to estimate national level malaria burden.

Modelling the impact of climate change on the distribution and abundance of tsetse in Northern Zimbabwe

Background

Climate change is predicted to impact the transmission dynamics of vector-borne diseases. Tsetse flies (Glossina) transmit species of Trypanosoma that cause human and animal African trypanosomiasis. A previous modelling study showed that temperature increases between 1990 and 2017 can explain the observed decline in abundance of tsetse at a single site in the Mana Pools National Park of Zimbabwe. Here, we apply a mechanistic model of tsetse population dynamics to predict how increases in temperature may have changed the distribution and relative abundance of Glossina pallidipes across northern Zimbabwe.

Methods

Local weather station temperature measurements were previously used to fit the mechanistic model to longitudinal G. pallidipes catch data. To extend the use of the model, we converted MODIS land surface temperature to air temperature, compared the converted temperatures with available weather station data to confirm they aligned, and then re-fitted the mechanistic model using G. pallidipes catch data and air temperature estimates. We projected this fitted model across northern Zimbabwe, using simulations at a 1 km × 1 km spatial resolution, between 2000 to 2016.

Results

We produced estimates of relative changes in G. pallidipes mortality, larviposition, emergence rates and abundance, for northern Zimbabwe. Our model predicts decreasing tsetse populations within low elevation areas in response to increasing temperature trends during 2000–2016. Conversely, we show that high elevation areas (> 1000 m above sea level), previously considered too cold to sustain tsetse, may now be climatically suitable.

Conclusions

To our knowledge, the results of this research represent the first regional-scale assessment of temperature related tsetse population dynamics, and the first high spatial-resolution estimates of this metric for northern Zimbabwe. Our results suggest that tsetse abundance may have declined across much of the Zambezi Valley in response to changing climatic conditions during the study period. Future research including empirical studies is planned to improve model accuracy and validate predictions for other field sites in Zimbabwe.

Exploratory Analysis of Urban Climate Using a Gap-Filled Landsat 8 Land Surface Temperature Data Set

The Landsat 8 satellites have retrieved land surface temperature (LST) resampled at a 30-m spatial resolution since 2013, but the urban climate studies frequently use a limited number of images due to the problems related to missing data over the city of interest. This paper endorses a procedure for building a long-term gap-free LST data set in an urban area using the high-resolution Landsat 8 imagery. The study is applied on 94 images available through 2013–2018 over Bucharest (Romania). The raw images containing between 1.1% and 58.4% missing LST data were filled in using the Data INterpolating Empirical Orthogonal Functions (DINEOF) algorithm implemented in the sinkr R packages. The resulting high-spatial-resolution gap-filled land surface temperature data set was used to explore the LST climatology over Bucharest (Romania) an urban area, at a monthly, seasonal, and annual scale. The performance of the gap-filling method was checked using a cross-validation procedure, and the results pledge for the development of an LST-based urban climatology.

A Simple Method for Converting 1-km Resolution Daily Clear-Sky LST into Real LST

Land-surface temperature (LST) plays a key role in the physical processes of surface energy and water balance from local through global scales. The widely used one kilometre resolution daily Moderate Resolution Imaging Spectroradiometer (MODIS) LST product has missing values due to the influence of clouds. Therefore, a large number of clear-sky LST reconstruction methods have been developed to obtain spatially continuous LST datasets. However, the clear-sky LST is a theoretical value that is often an overestimate of the real value. In fact, the real LST (also known as cloudy-sky LST) is more necessary and more widely used. The existing cloudy-sky LST algorithms are usually somewhat complicated, and the accuracy needs to be improved. It is necessary to convert the clear-sky LST obtained by the currently better-developed methods into cloudy-sky LST. We took the clear-sky LST, cloud-cover duration, downward shortwave radiation, albedo and normalized difference vegetation index (NDVI) as five independent variables and the real LST at the ground stations as the dependent variable to perform multiple linear regression. The mean absolute error (MAE) of the cloudy-sky LST retrieved by this method ranged from 3.5–3.9 K. We further analyzed different cases of the method, and the results suggested that this method has good flexibility. When we chose fewer independent variables, different clear-sky algorithms, or different regression tools, we also achieved good results. In addition, the method calculation process was relatively simple and can be applied to other research areas. This study preliminarily explored the influencing factors of the real LST and can provide a possible option for researchers who want to obtain cloudy-sky LST through clear-sky LST, that is, a convenient conversion method. This article lays the foundation for subsequent research in various fields that require real LST.

Comprehensive evaluation of a spatio-temporal gap filling algorithm: Using remotely sensed precipitation, LST and ET data

Temporal and spatial continuity of remote sensing data is flawed due to cloudiness, sensor malfunction or atmospheric pollution. Different methods have been presented to estimate missing values in remote sensing data. In this study, we evaluate the performance of a spatio-temporal gap filling algorithm developed by Weiss et al. (2014). This algorithm is interesting and worthy for further evaluation because it achieves high accuracy while maintaining the computational complexity considerably low. To conduct a comprehensive evaluation, we applied the algorithm to MODIS (Land Surface Temperature (LST) and evapotranspiration (ET)) and TRMM (precipitation) time series and investigate the effects of several factors including seasonality, variable type, gap size and surface characteristics through simulation scenarios. The performances were discussed using qualitative and quantitative assessments based on different simulation scenarios. A crucial finding of this study is a subtle structural deficiency of the algorithm. In particular, the algorithm outputs highly erroneous estimations when dealing with pixels with values mostly between zero and one. Such unexpected errors were observed in the seasonal assessment of land surface temperature estimations. In addition, according to the results of this study, the algorithm was sensitive to the variable type; however there was no correlation between the studied gap sizes and the error values. Among the three studied variables, LST and ET missing values were restored very accurately while estimations of precipitation missing values were more erroneous. The results also exhibit that in heterogeneous areas with complex topography, the errors of estimations were higher than homogeneous regions and areas with less complex topography. Based on the results, the algorithm should be used with caution in the discrete parameters like precipitation and area with abrupt variations. Furthermore, the design of the method may be refined for such datasets which include values with range between zero and one.

DATimeS: A machine learning time series GUI toolbox for gap-filling and vegetation phenology trends detection

Optical remotely sensed data are typically discontinuous, with missing values due to cloud cover. Consequently, gap-filling solutions are needed for accurate crop phenology characterization. The here presented Decomposition and Analysis of Time Series software (DATimeS) expands established time series interpolation methods with a diversity of advanced machine learning fitting algorithms (e.g., Gaussian Process Regression: GPR) particularly effective for the reconstruction of multiple-seasons vegetation temporal patterns. DATimeS is freely available as a powerful image time series software that generates cloud-free composite maps and captures seasonal vegetation dynamics from regular or irregular satellite time series. This work describes the main features of DATimeS, and provides a demonstration case using Sentinel-2 Leaf Area Index time series data over a Spanish site. GPR resulted as an optimum fitting algorithm with most accurate gap-filling performance and associated uncertainties. DATimeS further quantified LAI fluctuations among multiple crop seasons and provided phenological indicators for specific crop types.

Statistical Gap-Filling of SEVIRI Land Surface Temperature

A reliable and practically useable method for gap filling in hourly Spinning Enhanced Visible and Infrared Imager (SEVIRI LST) data using ERA5 Land Skin Temperature (ERA5ST) co-variate and additional easily accessible data (elevation, time, solar radiation info) is proposed. The suggested approach provides estimates to all weather conditions and it is based on a probabilistic model via modern regression models. We have tested two classes of regression models of different complexity and flexibility, namely multiple linear regression (MLR), and generalized additive model (GAM). This analysis uses as main input the hourly LST data set over Romania, through 2016 and 2017, extracted from MSG-SEVIRI images, which is an operational product of the Land Surface Analysis–Satellite Application Facility (LSA-SAF). The comparison between the estimated LST and the original LST values shows that GAM model, that takes into account the distance between missing LST locations and the nearest non-missing locations (GAM2), provides the best results, hence this was used to fill the gaps from the analyzed remote sensing product. Considering the fact that the best covariate (ERA5ST) has global coverage and it is available at high spatial resolution and temporal resolution, the proposed approach could be also used to perform the gap-filling of other existing LST remote sensing products.

Gap Fill of Land Surface Temperature and Reflectance Products in Landsat Analysis Ready Data

The recently released Landsat analysis ready data (ARD) over the United States provides the opportunity to investigate landscape dynamics using dense time series observations at 30-m resolution. However, the dataset often contains data gaps (or missing data) because of cloud contamination or data acquisition strategy, which result in different capabilities for seasonality modeling. We present a new algorithm that focuses on data gap filling using clear observations from orbit overlap regions. Multiple linear regression models were established for each pixel time series to estimate stable predictions and uncertainties. The model’s training data came from stratified random samples based on the time series similarity between the pixel and data from the overlap regions. The algorithm was first evaluated using four tiles (5000 × 5000 30-m pixels for each tile) from 2018 land surface temperature data (LST) in Atlanta, Georgia. The accuracy was assessed using randomly masked clear observations with an average Root Mean Square Error (RMSE) of 3.88 and an average bias of −0.37, which were comparable to the product accuracy. We also applied the method on ARD surface reflectance bands at Fairbanks, Alaska. The accuracy assessment suggested a majority RMSE of less than 0.04 and a bias of less than 0.0023. The gap-filled time series can be of help for reliable seasonal modeling and reducing artifacts related to data availability. This approach can also be applied to other datasets, vegetation indexes, or spectral reflectance bands of other sensors.

Vegetation Changes Attributable to Refugees in Africa Coincide with Agricultural Deforestation

The recent adoption of the Global Compact on Refugees formally recognizes not only the importance of supporting the nearly 26 million people who have sought asylum from conflict and persecution but also of easing the pressures on receiving areas and host countries. However, few countries may enforce the Compact out of concern over the economic or environmental repercussions of hosting refugees. We examine whether narratives of refugee-driven landscape change are empirically generalizable to continental Africa, which fosters 34% of all refugees. Estimates of the causal effects of the number of refugees-located in 493 camps distributed across 49 African countries-on vegetation from 2000 to 2016 are provided. Using a quasi-experimental design, we find refugees bear a small increase in vegetation condition while contributing to increased deforestation. Such a combination is mainly explained not by land clearance and massive biomass extraction but by agricultural expansion in refugee-hosting areas. A one percent increase in the number of refugees amplifies the transition from dominant forested areas to cropland by 1.4 percentage points. These findings suggest that changes in vegetation condition may ensue with the elevation of population-based constraints on food security.

Filling Gaps of Monthly Terra/MODIS Daytime Land Surface Temperature Using Discrete Cosine Transform Method

Land surface temperature (LST) is a key parameter in geophysical fields. The Moderate Resolution Imaging Spectroradiometer (MODIS) onboard Terra provides an accurate LST dataset with global coverage and monthly series, but the monthly MODIS LST data are often obscured by clouds and other atmospheric disturbances and consequently exhibit significant data gaps at a global scale, resulting in a difficult interpretation of LST trends and climatological characteristics. In this paper, an effective and fast LST reconstruction method to fill data gaps in monthly MODIS LST is presented. The proposal combines the Discrete Cosine Transform (DCT) and the Penalized Least Square approach (PLS) together with the Generalized Cross-Validation (GCV) criterion. It depends only on the spatial high-frequency information from original LST estimates and allows a fast and automatic filling process without the help of any other ancillary data. To analyze its performance, the method is applied to fill data gaps on three continents with synthetic random missing values introduced as validation sets. The statistical evaluation shows that this method is capable of filling a large number of missing values in MODIS LST datasets with very high accuracy. In addition, the trend differences between the original LST and reconstructed LST have assessed the significance by computing 95% confidence intervals for a time series of trend differences is examined. Simulated experiments show that data gaps with large missing counts lead to significant differences in trend patterns and the patterns on validation sets are well estimated by this method, which confirms that the filling process of MODIS LST is necessary and favorable results can be produced for substantial data gaps by the DCT-PLS method.

Validation of Earth Observation Time-Series: A Review for Large-Area and Temporally Dense Land Surface Products

Large-area remote sensing time-series offer unique features for the extensive investigation of our environment. Since various error sources in the acquisition chain of datasets exist, only properly validated results can be of value for research and downstream decision processes. This review presents an overview of validation approaches concerning temporally dense time-series of land surface geo-information products that cover the continental to global scale. Categorization according to utilized validation data revealed that product intercomparisons and comparison to reference data are the conventional validation methods. The reviewed studies are mainly based on optical sensors and orientated towards global coverage, with vegetation-related variables as the focus. Trends indicate an increase in remote sensing-based studies that feature long-term datasets of land surface variables. The hereby corresponding validation efforts show only minor methodological diversification in the past two decades. To sustain comprehensive and standardized validation efforts, the provision of spatiotemporally dense validation data in order to estimate actual differences between measurement and the true state has to be maintained. The promotion of novel approaches can, on the other hand, prove beneficial for various downstream applications, although typically only theoretical uncertainties are provided.

Reconstructing One Kilometre Resolution Daily Clear-Sky LST for China’s Landmass Using the BME Method

The land surface temperature (LST) is a key parameter used to characterize the interaction between land and the atmosphere. Therefore, obtaining highly accurate, spatially consistent and temporally continuous LSTs in large areas is the basis of many studies. The Moderate Resolution Imaging Spectroradiometer (MODIS) LST product is commonly used to achieve this. However, it has many missing values caused by clouds and other factors. The current gap-filling methods need to be improved when applied to large areas. In this study, we used the Bayesian maximum entropy (BME) method, which considers spatial and temporal correlation, and takes multiple regression results of the Normalized Difference Vegetation Index (NDVI), Digital Elevation Model (DEM), longitude and latitude as soft data to reconstruct space-complete daily clear-sky LSTs with a 1 km resolution for China’s landmass in 2015. The average Root Mean Square Error (RMSE) of this method was 1.6 K for the daytime and 1.2 K for the nighttime when we simultaneously covered more than 10,000 verification points, including blocks that were continuous in space, and the average RMSE of a single discrete verification point for 365 days was 0.4 K for the daytime and 0.3 K for the nighttime when we covered four discrete points. Urban and snow land cover types have a higher accuracy than forests and grasslands, and the accuracy is higher in winter than in summer. The high accuracy and great ability of this method to capture extreme values in urban areas can help improve urban heat island research. This method can also be extended to other study areas, other time periods, and the estimation of other geographical attribute values. How to effectively convert clear-sky LST into real LST requires further research.

Mapping Geospatial Processes Affecting the Environmental Fate of Agricultural Pesticides in Africa

The application of agricultural pesticides in Africa can have negative effects on human health and the environment. The aim of this study was to identify African environments that are vulnerable to the accumulation of pesticides by mapping geospatial processes affecting pesticide fate. The study modelled processes associated with the environmental fate of agricultural pesticides using publicly available geospatial datasets. Key geospatial processes affecting the environmental fate of agricultural pesticides were selected after a review of pesticide fate models and maps for leaching, surface runoff, sedimentation, soil storage and filtering capacity, and volatilization were created. The potential and limitations of these maps are discussed. We then compiled a database of studies that measured pesticide residues in Africa. The database contains 10,076 observations, but only a limited number of observations remained when a standard dataset for one compound was extracted for validation. Despite the need for more in-situ data on pesticide residues and application, this study provides a first spatial overview of key processes affecting pesticide fate that can be used to identify areas potentially vulnerable to pesticide accumulation.

Scavenging in the Anthropocene: Human impact drives vertebrate scavenger species richness at a global scale

Understanding the distribution of biodiversity across the Earth is one of the most challenging questions in biology. Much research has been directed at explaining the species latitudinal pattern showing that communities are richer in tropical areas; however, despite decades of research, a general consensus has not yet emerged. In addition, global biodiversity patterns are being rapidly altered by human activities. Here, we aim to describe large-scale patterns of species richness and diversity in terrestrial vertebrate scavenger (carrion-consuming) assemblages, which provide key ecosystem functions and services. We used a worldwide dataset comprising 43 sites, where vertebrate scavenger assemblages were identified using 2,485 carcasses monitored between 1991 and 2018. First, we evaluated how scavenger richness (number of species) and diversity (Shannon diversity index) varied among seasons (cold vs. warm, wet vs. dry). Then, we studied the potential effects of human impact and a set of macroecological variables related to climatic conditions on the scavenger assemblages. Vertebrate scavenger richness ranged from species-poor to species rich assemblages (4-30 species). Both scavenger richness and diversity also showed some seasonal variation. However, in general, climatic variables did not drive latitudinal patterns, as scavenger richness and diversity were not affected by temperature or rainfall. Rainfall seasonality slightly increased the number of species in the community, but its effect was weak. Instead, the human impact index included in our study was the main predictor of scavenger richness. Scavenger assemblages in highly human-impacted areas sustained the smallest number of scavenger species, suggesting human activity may be overriding other macroecological processes in shaping scavenger communities. Our results highlight the effect of human impact at a global scale. As species-rich assemblages tend to be more functional, we warn about possible reductions in ecosystem functions and the services provided by scavengers in human-dominated landscapes in the Anthropocene.

Quantifying the impacts of dams on riverine hydrology under non-stationary conditions using incomplete data and Gaussian copula models

Across the world, the assessment of environmental impacts attributable to infrastructure and development projects often require a comparison between observed post-impact outcomes with what "would have happened" in the absence of the impact (i.e., the counterfactual). Environmental impact assessment (EIA) methods traditionally determine the counterfactual based on strong assumptions of stationarity (e.g., using before and after comparisons) and can be particularly challenging to use in the context of substantial data gaps, a vexing problem when combining several time-series data from different sources. Here we propose and test a widely applicable statistical approach for quantifying environmental impacts that avoids the stationarity assumption and circumvents issues associated with data gaps. Specifically, we used a Gaussian Copula (GC) model to assess the hydrological impacts of the Tucuruí dam on the Tocantins River in the Brazilian Amazon. Using multi-source water level and climate data, GC predictions of pre-dam hydrology for the validation period were excellent (Nash-Sutcliffe coefficients of 0.83 to 0.98 and 93-96% of observations within the 95% predictive intervals). In the post-dam period, the river had higher dry-season water levels both upstream and downstream relative to the predicted counterfactual, and the timing and duration of wet-season drawdown was delayed and extended, substantially altering the flood pulse. These impacts were evident as far as 176 km away from the dam, highlighting widespread hydrological impacts. The GC model outperformed standard multiple regression models in representing predictive uncertainty while also avoiding the stationarity assumption and circumventing the issue of sparse and incomplete data. We thus believe the GC approach has wide utility for integrating disparate time-series data to quantify the impacts of dams and other anthropogenic phenomena on riverine hydrology globally.

A Decision Tree Approach for Spatially Interpolating Missing Land Cover Data and Classifying Satellite Images

Sustainable Development Goals (SDGs) are a set of priorities the United Nations and World Bank have set for countries to reach in order to improve quality of life and environment globally by 2030. Free satellite images have been identified as a key resource that can be used to produce official statistics and analysis to measure progress towards SDGs, especially those that are concerned with the physical environment, such as forest, water, and crops. Satellite images can often be unusable due to missing data from cloud cover, particularly in tropical areas where the deforestation rates are high. There are existing methods for filling in image gaps; however, these are often computationally expensive in image classification or not effective at pixel scale. To address this, we use two machine learning methods—gradient boosted machine and random forest algorithms—to classify the observed and simulated ‘missing’ pixels in satellite images as either grassland or woodland. We also predict a continuous biophysical variable, Foliage Projective Cover (FPC), which was derived from satellite images, and perform accurate binary classification and prediction using only the latitude and longitude of the pixels. We compare the performance of these methods against each other and inverse distance weighted interpolation, which is a well-established spatial interpolation method. We find both of the machine learning methods, particularly random forest, perform fast and accurate classifications of both observed and missing pixels, with up to 0.90 accuracy for the binary classification of pixels as grassland or woodland. The results show that the random forest method is more accurate than inverse distance weighted interpolation and gradient boosted machine for prediction of FPC for observed and missing data. Based on the case study results from a sub-tropical site in Australia, we show that our approach provides an efficient alternative for interpolating images and performing land cover classifications.

Large-Scale Crop Mapping Based on Machine Learning and Parallel Computation with Grids

Large-scale crop mapping provides important information in agricultural applications. However, it is a challenging task due to the inconsistent availability of remote sensing data caused by the irregular time series and limited coverage of the images, together with the low spatial resolution of the classification results. In this study, we proposed a new efficient method based on grids to address the inconsistent availability of the high-medium resolution images for large-scale crop classification. First, we proposed a method to block the remote sensing data into grids to solve the problem of temporal inconsistency. Then, a parallel computing technique was introduced to improve the calculation efficiency on the grid scale. Experiments were designed to evaluate the applicability of this method for different high-medium spatial resolution remote sensing images and different machine learning algorithms and to compare the results with the widely used nonparallel method. The computational experiments showed that the proposed method was successful at identifying large-scale crop distribution using common high-medium resolution remote sensing images (GF-1 WFV images and Sentinel-2) and common machine learning classifiers (the random forest algorithm and support vector machine). Finally, we mapped the croplands in Heilongjiang Province in 2015, 2016, 2017, which used a random forest classifier with the time series GF-1 WFV images spectral features, the enhanced vegetation index (EVI) and normalized difference water index (NDWI). Ultimately, the accuracy was assessed using a confusion matrix. The results showed that the classification accuracy reached 88%, 82%, and 85% in 2015, 2016, and 2017, respectively. In addition, with the help of parallel computing, the calculation speed was significantly improved by at least seven-fold. This indicates that using the grid framework to block the data for classification is feasible for crop mapping in large areas and has great application potential in the future.

Comparing Remotely-Sensed Surface Energy Balance Evapotranspiration Estimates in Heterogeneous and Data-Limited Regions: A Case Study of Tanzania’s Kilombero Valley

Evapotranspiration (ET) plays a crucial role in integrated water resources planning, development and management, especially in tropical and arid regions. Determining ET is not straightforward due to the heterogeneity and complexity found in real-world hydrological basins. This situation is often compounded in regions with limited hydro-meteorological data that are facing rapid development of irrigated agriculture. Remote sensing (RS) techniques have proven useful in this regard. In this study, we compared the daily actual ET estimates derived from 3 remotely-sensed surface energy balance (SEB) models, namely, the Surface Energy Balance Algorithm for Land (SEBAL) model, the Operational Simplified Surface Energy Balance (SSEBop) model, and the Simplified Surface Balance Index (S-SEBI) model. These products were generated using the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery for a total of 44 satellite overpasses in 2005, 2010, and 2015 in the heterogeneous, highly-utilized, rapidly-developing and data-limited Kilombero Valley (KV) river basin in Tanzania, eastern Africa. Our results revealed that the SEBAL model had a relatively high ET compared to other models and the SSEBop model had relatively low ET compared to the other models. In addition, we found that the S-SEBI model had a statistically similar ET as the ensemble mean of all models. Further comparison of SEB models’ ET estimates across different land cover classes and different spatial scales revealed that almost all models’ ET estimates were statistically comparable (based on the Wilcoxon’s test and the Levene’s test at a 95% confidence level), which implies fidelity between and reliability of the ET estimates. Moreover, all SEB models managed to capture the two spatially-distinct ET regimes in KV: the stable/permanent ET regime on the mountainous parts of the KV and the seasonally varied ET over the floodplain which contains a Ramsar site (Kilombero Valley Floodplain). Our results have the potential to be used in hydrological modelling to explore and develop integrated water resources management in the valley. We believe that our approach can be applied elsewhere in the world especially where observed meteorological variables are limited.

Near real-time vegetation anomaly detection with MODIS NDVI: Timeliness vs. accuracy and effect of anomaly computation options

For food crises early warning purposes, coarse spatial resolution NDVI data are widely used to monitor vegetation conditions in near real-time (NRT). Different types of NDVI anomalies are typically employed to assess the current state of crops and rangelands as compared to previous years. Timeliness and accuracy of such anomalies are critical factors to an effective monitoring. Temporal smoothing can efficiently reduce noise and cloud contamination in the time series of historical observations, where data points are available before and after each observation to be smoothed. With NRT data, smoothing methods are adapted to cope with the unbalanced availability of data before and after the most recent data points. These NRT approaches provide successive updates of the estimation of the same data point as more observations become available. Anomalies compare the current NDVI value with some statistics (e.g. indicators of central tendency and dispersion) extracted from the historical archive of observations. With multiple updates of the same datasets being available, two options can be selected to compute anomalies, i.e. using the same update level for the NRT data and the statistics or using the most reliable update for the latter. In this study we assess the accuracy of three commonly employed 1 km MODIS NDVI anomalies (standard scores, non-exceedance probability and vegetation condition index) with respect to (1) delay with which they become available and (2) option selected for their computation. We show that a large estimation error affects the earlier estimates and that this error is efficiently reduced in subsequent updates. In addition, with regards to the preferable option to compute anomalies, we empirically observe that it depends on the type of application (e.g. averaging anomalies value over an area of interest vs. detecting "drought" conditions by setting a threshold on the anomaly value) and the employed anomaly type. Finally, we map the spatial pattern in the magnitude of NRT anomaly estimation errors over the globe and relate it to average cloudiness.

Satellite Remote Sensing of Surface Urban Heat Islands: Progress, Challenges, and Perspectives

The surface urban heat island (SUHI), which represents the difference of land surface temperature (LST) in urban relativity to neighboring non-urban surfaces, is usually measured using satellite LST data. Over the last few decades, advancements of remote sensing along with spatial science have considerably increased the number and quality of SUHI studies that form the major body of the urban heat island (UHI) literature. This paper provides a systematic review of satellite-based SUHI studies, from their origin in 1972 to the present. We find an exponentially increasing trend of SUHI research since 2005, with clear preferences for geographic areas, time of day, seasons, research foci, and platforms/sensors. The most frequently studied region and time period of research are China and summer daytime, respectively. Nearly two-thirds of the studies focus on the SUHI/LST variability at a local scale. The Landsat Thematic Mapper (TM)/Enhanced Thematic Mapper (ETM+)/Thermal Infrared Sensor (TIRS) and Terra/Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) are the two most commonly-used satellite sensors and account for about 78% of the total publications. We systematically reviewed the main satellite/sensors, methods, key findings, and challenges of the SUHI research. Previous studies confirm that the large spatial (local to global scales) and temporal (diurnal, seasonal, and inter-annual) variations of SUHI are contributed by a variety of factors such as impervious surface area, vegetation cover, landscape structure, albedo, and climate. However, applications of SUHI research are largely impeded by a series of data and methodological limitations. Lastly, we propose key potential directions and opportunities for future efforts. Besides improving the quality and quantity of LST data, more attention should be focused on understudied regions/cities, methods to examine SUHI intensity, inter-annual variability and long-term trends of SUHI, scaling issues of SUHI, the relationship between surface and subsurface UHIs, and the integration of remote sensing with field observations and numeric modeling.

A Case Study Competition Among Methods for Analyzing Large Spatial Data

The Gaussian process is an indispensable tool for spatial data analysts. The onset of the "big data" era, however, has lead to the traditional Gaussian process being computationally infeasible for modern spatial data. As such, various alternatives to the full Gaussian process that are more amenable to handling big spatial data have been proposed. These modern methods often exploit low-rank structures and/or multi-core and multi-threaded computing environments to facilitate computation. This study provides, first, an introductory overview of several methods for analyzing large spatial data. Second, this study describes the results of a predictive competition among the described methods as implemented by different groups with strong expertise in the methodology. Specifically, each research group was provided with two training datasets (one simulated and one observed) along with a set of prediction locations. Each group then wrote their own implementation of their method to produce predictions at the given location and each was subsequently run on a common computing environment. The methods were then compared in terms of various predictive diagnostics. Supplementary materials regarding implementation details of the methods and code are available for this article online.

ELECTRONIC SUPPLEMENTARY MATERIAL

Supplementary materials for this article are available at 10.1007/s13253-018-00348-w.

Spatial mapping with Gaussian processes and nonstationary Fourier features

The use of covariance kernels is ubiquitous in the field of spatial statistics. Kernels allow data to be mapped into high-dimensional feature spaces and can thus extend simple linear additive methods to nonlinear methods with higher order interactions. However, until recently, there has been a strong reliance on a limited class of stationary kernels such as the Matérn or squared exponential, limiting the expressiveness of these modelling approaches. Recent machine learning research has focused on spectral representations to model arbitrary stationary kernels and introduced more general representations that include classes of nonstationary kernels. In this paper, we exploit the connections between Fourier feature representations, Gaussian processes and neural networks to generalise previous approaches and develop a simple and efficient framework to learn arbitrarily complex nonstationary kernel functions directly from the data, while taking care to avoid overfitting using state-of-the-art methods from deep learning. We highlight the very broad array of kernel classes that could be created within this framework. We apply this to a time series dataset and a remote sensing problem involving land surface temperature in Eastern Africa. We show that without increasing the computational or storage complexity, nonstationary kernels can be used to improve generalisation performance and provide more interpretable results.

Determining environmental and anthropogenic factors which explain the global distribution of Aedes aegypti and Ae. albopictus

Background

Responsible for considerable global human morbidity and mortality, Aedes aegypti and Ae. albopictus are the primary vectors of several important human diseases, including dengue and yellow fever. Although numerous variables that affect mosquito survival and reproduction have been recorded at the local and regional scales, many remain untested at the global level, potentially confounding mapping efforts to date.

Methods

We develop a modelling ensemble of boosted regression trees and maximum entropy models using sets of variables previously untested at the global level to examine their performance in predicting the global distribution of these two vectors. The results show that accessibility, absolute humidity and annual minimum temperature are consistently the strongest predictors of mosquito presence. Both vectors are similar in their response to accessibility and humidity, but exhibit individual profiles for temperature. Their mapped ranges are therefore similar except at peripheral latitudes, where the range of Ae. albopictus extends further, a finding consistent with ongoing trapping studies. We show that variables previously identified as being relevant, including maximum and mean temperatures, enhanced vegetation index, relative humidity and population density, are comparatively weak performers.

Results

The variables identified represent three key biological mechanisms. Cold tolerance is a critical biological parameter, controlling both species' distribution northwards, and to a lesser degree for Ae. albopictus which has consequent greater inland suitability in North America, Europe and East Asia. Absolute humidity restricts the distribution of both vectors from drier areas, where moisture availability is very low, and increases their suitability in coastal areas. The latter is exacerbated by accessibility with increased likelihood of vector importation due to greater potential for human and trade movement.

Conclusion

Accessibility, absolute humidity and annual minimum temperatures were the strongest and most robust global predictors of Ae. aegypti and Ae. albopictus presence, which should be considered in control efforts and future distribution projections.

Genomic and epidemiological monitoring of yellow fever virus transmission potential

The yellow fever virus (YFV) epidemic in Brazil is the largest in decades. The recent discovery of YFV in Brazilian Aedes species mosquitos highlights a need to monitor the risk of reestablishment of urban YFV transmission in the Americas. We use a suite of epidemiological, spatial, and genomic approaches to characterize YFV transmission. We show that the age and sex distribution of human cases is characteristic of sylvatic transmission. Analysis of YFV cases combined with genomes generated locally reveals an early phase of sylvatic YFV transmission and spatial expansion toward previously YFV-free areas, followed by a rise in viral spillover to humans in late 2016. Our results establish a framework for monitoring YFV transmission in real time that will contribute to a global strategy to eliminate future YFV epidemics.

Spatio-temporal mapping of Madagascar's Malaria Indicator Survey results to assess Plasmodium falciparum endemicity trends between 2011 and 2016

BACKGROUND

Reliable measures of disease burden over time are necessary to evaluate the impact of interventions and assess sub-national trends in the distribution of infection. Three Malaria Indicator Surveys (MISs) have been conducted in Madagascar since 2011. They provide a valuable resource to assess changes in burden that is complementary to the country's routine case reporting system.

METHODS

A Bayesian geostatistical spatio-temporal model was developed in an integrated nested Laplace approximation framework to map the prevalence of Plasmodium falciparum malaria infection among children from 6 to 59 months in age across Madagascar for 2011, 2013 and 2016 based on the MIS datasets. The model was informed by a suite of environmental and socio-demographic covariates known to influence infection prevalence. Spatio-temporal trends were quantified across the country.

RESULTS

Despite a relatively small decrease between 2013 and 2016, the prevalence of malaria infection has increased substantially in all areas of Madagascar since 2011. In 2011, almost half (42.3%) of the country's population lived in areas of very low malaria risk (<1% parasite prevalence), but by 2016, this had dropped to only 26.7% of the population. Meanwhile, the population in high transmission areas (prevalence >20%) increased from only 2.2% in 2011 to 9.2% in 2016. A comparison of the model-based estimates with the raw MIS results indicates there was an underestimation of the situation in 2016, since the raw figures likely associated with survey timings were delayed until after the peak transmission season.

CONCLUSIONS

Malaria remains an important health problem in Madagascar. The monthly and annual prevalence maps developed here provide a way to evaluate the magnitude of change over time, taking into account variability in survey input data. These methods can contribute to monitoring sub-national trends of malaria prevalence in Madagascar as the country aims for geographically progressive elimination.

Large-Area Gap Filling of Landsat Reflectance Time Series by Spectral-Angle-Mapper Based Spatio-Temporal Similarity (SAMSTS)

Landsat time series commonly contain missing observations, i.e., gaps, due to the orbit and sensing geometry, data acquisition strategy, and cloud contamination. A spectral-angle-mapper (SAM) based spatio-temporal similarity (SAMSTS) gap-filling algorithm is presented that is designed to fill small and large area gaps in Landsat data, using one year or less of data and without using other satellite data. Each gap pixel is filled by an alternative similar pixel that is located in a non-missing region of the image. The alternative similar pixel locations are identified by comparison of reflectance time series using a SAM metric revised to be adaptive to missing observations. A time series segmentation-and-clustering approach is used to increase the search efficiency. The SAMSTS algorithm is demonstrated using six months of Landsat 8 Operational Land Imager (OLI) reflectance time series over three 150 × 150 km (5000 × 5000 30 m pixels) areas in California, Minnesota and Kansas. The three areas contain different land cover types, especially crops that have different phenology and abrupt changes due to agricultural harvesting, which make gap filling challenging. Fillings on simulated gaps, which are equivalent to 36% of 5000 × 5000 images in each test area, are presented. The gap filling accuracy is assessed quantitatively, and the SAMSTS algorithm is shown to perform better than the simple closest temporal pixel substitution gap filling approach and the sinusoidal harmonic model-based gap filling approach. The SAMSTS algorithm provides gap-filled data with five-band reflective-wavelength root-mean-square differences less the 0.02, which is comparable to the OLI reflectance calibration accuracy.

Existing and potential infection risk zones of yellow fever worldwide: a modelling analysis

Background

Yellow fever cases are under-reported and the exact distribution of the disease is unknown. An effective vaccine is available but more information is needed about which populations within risk zones should be targeted to implement interventions. Substantial outbreaks of yellow fever in Angola, Democratic Republic of the Congo, and Brazil, coupled with the global expansion of the range of its main urban vector, Aedes aegypti, suggest that yellow fever has the propensity to spread further internationally. The aim of this study was to estimate the disease's contemporary distribution and potential for spread into new areas to help inform optimal control and prevention strategies.

Methods

We assembled 1155 geographical records of yellow fever virus infection in people from 1970 to 2016. We used a Poisson point process boosted regression tree model that explicitly incorporated environmental and biological explanatory covariates, vaccination coverage, and spatial variability in disease reporting rates to predict the relative risk of apparent yellow fever virus infection at a 5 × 5 km resolution across all risk zones (47 countries across the Americas and Africa). We also used the fitted model to predict the receptivity of areas outside at-risk zones to the introduction or reintroduction of yellow fever transmission. By use of previously published estimates of annual national case numbers, we used the model to map subnational variation in incidence of yellow fever across at-risk countries and to estimate the number of cases averted by vaccination worldwide.

Findings

Substantial international and subnational spatial variation exists in relative risk and incidence of yellow fever as well as varied success of vaccination in reducing incidence in several high-risk regions, including Brazil, Cameroon, and Togo. Areas with the highest predicted average annual case numbers include large parts of Nigeria, the Democratic Republic of the Congo, and South Sudan, where vaccination coverage in 2016 was estimated to be substantially less than the recommended threshold to prevent outbreaks. Overall, we estimated that vaccination coverage levels achieved by 2016 avert between 94 336 and 118 500 cases of yellow fever annually within risk zones, on the basis of conservative and optimistic vaccination scenarios. The areas outside at-risk regions with predicted high receptivity to yellow fever transmission (eg, parts of Malaysia, Indonesia, and Thailand) were less extensive than the distribution of the main urban vector, A aegypti, with low receptivity to yellow fever transmission in southern China, where A aegypti is known to occur.

Interpretation

Our results provide the evidence base for targeting vaccination campaigns within risk zones, as well as emphasising their high effectiveness. Our study highlights areas where public health authorities should be most vigilant for potential spread or importation events.

Funding

Bill & Melinda Gates Foundation.