Multi-source Approach Research Articles

A multi-source approach combining GEDI LiDAR, satellite data, and machine learning algorithms for estimating forest aboveground biomass on Google Earth engine platform

A multi-source approach combining GEDI LiDAR, satellite data, and machine learning algorithms for estimating forest aboveground biomass on Google Earth engine platform

Multi-method approach effectively identifies and engages children impacted by parental substance misuse into school-based prevention services.

Children living in households where parents or caregiving adults misuse substances face significant risk academically, socially, physically, and emotionally. An estimated 12% or more of U.S. children lived with an adult with past-year substance use problems in 2009-2014. Engaging this high-need, underserved population in targeted prevention services is a public health imperative, requiring children first be identified. School-based services reduce access barriers and promote equitable access, providing a model that can address the scope and significance of parental substance misuse (PSM) on children. A review of published studies for this population revealed a lack of information regarding identification strategies and their relative effectiveness. This article uses data from a 2010-2020 field-based evaluation to analyze the performance of the Kids Like Us (KLU) program's manualized approach to identifying and engaging elementary students impacted by PSM into its school-based program. KLU, a program of the Frederick County Health Department (Maryland, U.S.), is implemented in collaboration with public school counselors. KLU's multi-method approach achieves universal prevention outcomes while simultaneously providing parent, self, school counselor, and community referral options. Over the 10-year study, 537 students were identified with a notable 83% of referred students completing 75-100% of sessions offered. Parent referral, a strategy not included in any reviewed studies, contributed the highest percentage of referrals (44% in response to a student take-home letter alone), followed by self (18%), school counselor (13%), and other/a combination (24%). KLU engaged students across varying school settings, sizes, and socioeconomics. Chi-square analysis of gender and ethnicity found no significant differences. Referral outcome and school counselor data results highlight the benefits of KLU's multi-method, multi-source approach, including its ease of replication. KLU's approach holds promise for addressing the public health crisis of children impacted by PSM. Study results highlight the need for policy changes including that U.S. and world alcohol and drug reports assess parent status. Additional research with a larger, more ethnically diverse population is recommended to examine the relationship between family and student demographics and referral strategies and sources.

Spatial Analysis of Sumateran Elephant (Elephas maximus sumatranus) Movement in Giam Siak Kecil (GSK) Elephant Enclave, Riau

The Sumatran elephant is currently classified as critically endangered. A variety of conservation and conflict management initiatives have been implemented with the objective of reducing the intensity of conflict between elephants and humans. The use of GPS collars for monitoring purposes not only serves to reduce conflict but also allows for the continuous monitoring of Sumatran elephant movements. This study aims to analyze the spatial distribution and movement patterns of Sumatran elephants in GSK Elephant Pockets based on habitat factors, namely land cover class, slope, distance from rivers, and distance from roads. The research was conducted between December 2023 to June 2024 GSK Elephant Enclave, Riau. This study employed a multi-source approach to obtain data on the movement patterns of Sumatran elephants. These included data from the Sumatran Elephant GPS Collar, which recorded the animals' movements between 2021-2023. The movement characteristics analyzed include the length of the path traversed by the Sumatran elephant, Nearest neighbor classification was employed for the analysis of spatial distribution patterns. The result showed that spatial distribution pattern of Sumatran elephants in GSK Elephant Pockets 2021-2023 is clustered with an NNI of 0.47 (NNI < 1). The highest number of Sumatran elephant movements in 2023 based on habitat factors, namely land cover class, was in plantation forest, with 4,788 movements (82.23%). The next highest number of movements was in flat slope class, with 3,229 movements (55.45%). The distance from the river was 0–500 m, with 2,323 movements (39.89%), while the distance from the road was 0–500 m, with 5,390 movements (92.56%).

Multi-source tri-environmental conceptual framework for fire impact analysis

The summer of 2023 highlighted the growing threat of wildfires in isolated areas, with Maui Island in Hawaii experiencing one of its largest wildfires. Our study employs a multi-source approach, integrating high-resolution optical data, thermal infrared data, and demographic information to assess the environmental, built, and social impacts of this event. Our innovative tri-environmental framework reveals significant vegetation degradation, land cover change, and disproportionate effects on various demographic groups. The fire caused extensive damage, with residential properties incurring 77.6% of the total losses, equating to approximately $563.2 million. Social impacts were profound, particularly among females, children, and the elderly, with employment and commuting disruptions affecting both low- and high-income groups. The study highlights the effectiveness of combining dasymetric mapping with real-time satellite data to refine population distribution estimates in affected areas. Our findings are applicable beyond wildfires, offering valuable insights into disaster response and mitigation strategies across various natural hazards like floods and earthquakes.

Transfer learning in bridge monitoring: Laboratory study on domain adaptation for population-based SHM of multispan continuous girder bridges

The presence of sufficient labelled data associated to various environmental conditions and damage scenarios often represents a challenge for the applicability of supervised-learning methods when dealing with structural health monitoring of real-scale infrastructures. To address this problem, population-based structural health monitoring has been recently proposed as an attractive solution, with the goal to collect information from similar structures and transfer health-state labels across the population. This paper focusses on the use of a feature-based transfer learning method. A machine-learning model is trained with source labelled data in a transformed features space to afterwards classify the unlabelled target dataset of a different bridge. More specifically, a domain adaptation-based methodology proposing two possible strategies, a single-source or a multi-source approach, is described. Given the difficulties in validating these techniques on real and varied datasets from multiple bridges, this paper presents a physical benchmark for population-based structural health monitoring in civil-engineering applications. The transfer between different configurations of a laboratory-scale bridge model, subjected to multiple experimental tests under changing environmental conditions and to the same pseudo-damage scenarios, is investigated. The results of the experimental campaign demonstrate the possibility of effectively exchanging damage labels to perform novelty detection and damage classification across the population via domain adaptation, improve the identification of specific damage classes, as well as to increase model’s outcome using a multi-source approach, thus overcoming the limitations of conventional machine learning-based methods. Furthermore, this paper provides an open dataset with the physical benchmark-related data, allowing other researchers to test their own algorithms and address the various transfer learning challenges.

A multi-source approach to mapping habitat diversity: Comparison and combination of single-date hyperspectral and multi-date multispectral satellite imagery in a Mediterranean Natural Reserve

The increasing availability of spaceborne hyperspectral satellite imagery opens new opportunities for forest habitat mapping and monitoring, but the limitation of its generally low temporal resolution must be considered. In this study, we compare the ability of single-date PRISMA (PRecursore IperSpettrale della Missione Applicativa), the hyperspectral satellite from the Italian Space Agency, with that of both single-date and multi-date Sentinel-2 (S2) and PlanetScope (PS) to detect and correctly classify various EUNIS habitat types distributed over a relatively small spatial extent (6000 ha) in a natural reserve in Central Italy. The case study deals with multiple levels of spectral similarity, as the dominant canopy species of the target forest habitat classes belong to the same genus (Quercus spp., both deciduous and evergreen species) as well as of different taxa (Pinus and Fraxinus spp.). We performed a pixel-based classification with the Random Forest algorithm using a set of 28 spectral indices computed on PRISMA bands, 22 on S2, and 12 on PS. A Canopy Height Model (CHM) was also used as an input variable for the classification. Our results showed that PRISMA considerably outperforms the two multispectral satellites in single-date classifications, with an overall accuracy of 84 % compared to PlanetScope's 69 % and Sentinel-2's 72 %. Regarding the comparison between multi-date multispectral and single-date hyperspectral, 10-fold cross-validation results revealed that S2 achieves an out-of-bag error rate of approximately 16 %, while PRISMA achieves 17 % and PS 19 %. This demonstrates that a combination of spectral indices calculated during the growing season can capture phenological or physiological differences among the target species, which consequently results in a significant improvement in the classification accuracy of the multispectral sensors. Ultimately, classification results from all three sensors were combined to create probability maps for each forest class, identifying areas classified with a higher degree of certainty by each satellite tested and potentially contributing to forest management by defining areas with varying conservation levels.

A Methodologic Approach to Study Large and Complex Landslides: An Application in Central Apennines

The evaluation of landslide hazards in seismic areas is based on a deterministic analysis, which is unable to account for various uncertainties in the analysis process. This paper focuses on the probabilistic local seismic hazard analysis and extends the results to the landslide hazard analysis to consider both the uncertainties of the ground deformations and the strengths. The work studies the areas between Nibbiano and Sant’Erasmo hamlets in the Camerino municipality located in central Italy, where all constructions present evidence of damage caused by both the seismic sequence of 2016–2017 and the slope instability. An exhaustive geological and geophysical investigation has clarified the geological, geomorphological, and hydrogeological characteristics of the area, enabling a new characterization of material stress-strain behaviour. The study reveals that the low stiffness of the debris covers, and their fair degree of permeability contribute to potential instability scenarios triggered by both intense rainfall and the effects of strong earthquakes. The goal was to utilize the results to support local urban planning because in-depth knowledge of the possible evolutionary scenarios of the slopes is fundamental to the management of the degree of danger for structures, especially for people. Moreover, it was shown once again how a multi-source approach, with different investigation techniques, cannot be ignored for the study of the evolution of complex landslides.

Sustainable supply chain, dynamic capabilities, eco‐innovation, and environmental performance in an emerging economy

AbstractOrganizations can reduce costs and achieve better performance by utilizing supply chain management. However, they face enormous challenges in building eco‐friendly supply chains for harmony with nature, especially in emerging economies. Therefore, the current study seeks to determine the direct impact of sustainable supply chain management practices (SSCM) on environmental performance in Pakistani industries. The mediating role of eco‐innovation and sustainable, dynamic capabilities and the moderating role of institutional support were also tested. The data from 309 key informants were collected using a multisource approach and analyzed via structural equation modeling. The results indicated that SSCM and eco‐innovation positively impact environmental performance, but dynamic capabilities have negatively influenced environmental performance. Additionally, dynamic capabilities and eco‐innovation mediate the relationship between SSCM and eco‐innovation, and institutional support moderates the relationships positively. These results will enhance the theoretical base in supply chain management literature by exploring the role of SSCM in environmental performance. For practical implications, these results help managers in using institutional support to achieve environmental goals.

From regional signal to alert in addictovigilance

France has several monitoring systems that form the foundation of its health safety surveillance. This system is designed for the early detection of signals and their swift into an action system that enables timely, appropriate, and effective interventions to protect public health. These signals are considered alerts when sufficiently validated after an initial risk assessment and if they represent a potential threat to public health, necessitating an appropriate response. The French Addictovigilance Network (FAN) was established specifically to address complications associated with the use of psychoactive substances (PAS), both medicinal and non-medicinal, excluding alcohol and tobacco. This system, unique in Europe, provides information on one of the most complex vigilance systems due to its scope, and the risks associated with it are becoming an increasingly significant public health issue. FAN is at the forefront of identifying potential signals. It has been working closely with regional partners since 1990, particularly through a multi-source approach that combines data from various sources, in collaboration with local professionals. After outlining the missions of the addictovigilance centres and the general principles of signal detection in addictovigilance, this article will present recent examples involving similar products but different identifications or modes of consumption: synthetic cannabinoids, nitazene, pregabalin, methylphenidate. These examples demonstrate that while there are regional disparities, the risk of occurrence in other French regions should not be overlooked, and information should be rapidly communicated at the national level for preventive action.

Inaccurate Equals Bad? The Effect of Motivational Performance Appraisal Politics on Organizational Citizenship Behavior

Abstract: Performance appraisal politics are prevalent in organizations and consist of personal bias appraisal politics and motivational performance appraisal politics (MPAP). Although personal bias performance appraisal politics is deleterious, MPAP is purported to bring benefits to organizations. However, few empirical studies have proved the benefits of MPAP. Based on social exchange theory, this study posited that MPAP is positively related to organizational citizenship behavior (OCB) via perceived organizational support and that core self-evaluation weakens the positive relationship between MPAP and perceived organizational support and attenuates the relationship between MPAP and OCB via perceived organizational support. Using data from a sample of 246 employees collected through a two-wave and multisource approach and structural equation model, this study found that all hypotheses were supported.

Development and Evaluation of a Multisource Approach to Extend Mortality Follow-Up for Older Adults With Advanced Cancer Enrolled in Randomized Trials.

Mortality data can complement primary end points from cancer clinical trials. Yet, identifying deaths after trial completion is challenging, as timely and comprehensive vital status data are unavailable in the United States. We developed and evaluated a multisource approach to capture death data after clinical trial completion. Individuals age 70 years and older with incurable solid tumors or lymphoma and ≥1 aging-related condition were enrolled from October 2014 to March 2019 (ClinicalTrials.gov identifier: NCT02107443 and NCT02054741). Participants provided consent to link trial information to external sources. We developed a stepped approach for extended death capture using (1) active trial follow-up up to 1 year, (2) linkage to the National Death Index (NDI), and (3) obituary searches, thus generating a 5-year survival curve. In a random sample of 50 participants who died during trial follow-up, we estimated sensitivity of death data using NDI and obituary sources and computed survival times by data source. The two trials enrolled 1,169 participants; mean age was 76 years; 46% were female; and gastrointestinal cancer (30%) and lung cancer (26%) were the most common cancer types. Across data sources, maximum follow-up was >7 years; 5-year survival was 18%. In total, there were 841 deaths: 603 identified during trial follow-up; 199 from the NDI; and 39 from obituary searches. The sensitivity for death capture was 92% for the NDI and 94% for the obituary searches compared with the trial data, and computed survival times were similar across data sources. Extending clinical trial mortality follow-up through linkage with external data sources was feasible and accurate. Future cancer clinical trials should collect necessary consent and patient identifiers for vital status linkages that can enhance understanding of longer-term outcomes.

Effectiveness of an Immersive VR System for Construction Site Planning Education

Virtual reality (VR) has been widely used in education and training, but few studies have used it in education on construction site planning. Accordingly, this study developed an immersive VR-based education system for construction site planning and evaluated it in classroom-based tests with 114 students. Statistical analyses revealed that the VR system improved students' learning outcomes compared when conventional paper-based learning. Specifically, students with lower pretest subject knowledge achieved similar scores to top-ranked students after undergoing VR education. Moreover, VR education was more effective for learning concepts than it was for memorization. A single-source teaching approach was also more effective in improving learning outcomes than a multisource approach for VR learning. In summary, this study investigated three previously unexplored areas: targeting the ideal recipients for VR education and training, determining the appropriate problem types for VR education and training, and identifying the effectiveness of VR education and training for construction site planning.

Psychological Safety and Work Design as Mediators of Supervisors' Dark Triad Traits Impact on Nurses' Task Performance.

This study investigates how nurse supervisors' Dark Triad personality traits (Machiavellianism, narcissism, psychopathy) influence nurses' task performance, mediated by perceptions of enriched work design (autonomy, task variety, social support, safe work conditions, feedback quality) and psychological safety. A multisource approach was used to collect data from 256 manager-nurse dyads across various healthcare settings. Nurses completed surveys assessing their work design and psychological safety. Managers completed a self-assessment of Dark Triad traits and rated their nurse subordinates' task performance. Confirmatory factor analysis and structural equation modeling (SEM) were used for analysis. Supervisors' Dark Triad traits core component impacted nurses' task performance indirectly, mediated by psychological safety and nurses' perceptions of their enriched work design. Psychopathic traits revealed a significant direct negative effect on nurses' performance, while other Dark Triad traits did not show direct effects. This study sheds light on key factors influencing nurses' performance, offering insights for healthcare organizations aiming to optimize work environments and improve team effectiveness.

Enhancing System Reliability and Battery Longevity through Integrated Energy Sources and Algorithmic Optimization

This research examines the enhancement of low power electronic systems, such as low power wide area networks (LPWANs), by incorporating a mix of renewable energy sources like solar and wind in China. It addresses the challenges posed by the variability of these sources due to climatic and temporal factors by merging diverse energy harvesting (EH) methods to bolster power stability and availability. The study assesses the most effective combination of solar, wind, and rain EH for steady power supply. A novel algorithm is introduced for energy input management based on real‐time resource availability, with the goal of prolonging battery life by avoiding complete charge or discharge cycles. The feasibility of this multisource approach in possibly reducing the need for extra energy storage is evaluated. The investigation includes simulations using Chinese weather data to explore different combinations of energy sources. A real‐life system utilizing solar and wind energy, guided by the developed algorithm, was also implemented for empirical comparison. Findings suggest that solar and wind energies have a higher power yield than rain harvesting, with actual wind energy collection often falling short of simulated forecasts. The solar–wind energy mix achieves a 99% system availability and facilitates a more compact design owing to their high power density. Although the input‐switching algorithm lessens the frequency of complete battery drainage, a small degree of energy storage is still essential for maintaining system reliability. This integrated EH method, compared to single‐source options, allows for a smaller energy storage requirement. Moreover, the algorithm limits battery charging to 80%, significantly enhancing battery lifespan.

Mapping local-scale working population and daytime population densities using points-of-interest and nighttime light satellite imageries

ABSTRACT In this study, we proposed a multi-source approach for mapping local-scale population density of England. Specifically, we mapped both the working and daytime population densities by integrating the multi-source data such as residential population density, point-of-interest density, point-of-interest category mix, and nighttime light intensity. It is demonstrated that combining remote sensing and social sensing data provides a plausible way to map annual working or daytime population densities. In this paper, we trained models with England-wide data and subsequently tested these models with Wales-wide data. In addition, we further tested the models with England-wide data at a higher level of spatial granularity. Particularly, the random forest and convolutional neural network models were adopted to map population density. The estimated results and validation suggest that the three built models have high prediction accuracies at the local authority district level. It is shown that the convolutional neural network models have the greatest prediction accuracies at the local authority district level though they are most time-consuming. The models trained with the data at the local authority district level are less appropriately applicable to test data at a higher level of spatial granularity. The proposed multi-source approach performs well in mapping local-scale population density. It indicates that combining remote sensing and social sensing data is advantageous to mapping socioeconomic variables.

Rates of sudden cardiac death in the young down by 50%: insights from a 20-year nationwide study in Denmark

Abstract Background/Introduction Sudden cardiac death in the young (SCDY) constitutes an important public health problem and many years lost to society. Several initiatives have been implemented to reduce the rate of SCDY, e.g. improved resuscitation. Declining rates of overall and cardiovascular disease mortality have been well-documented; however, reliable information on the contemporary incidence of SCD is sparse. Purpose This study aimed to evaluate the temporal trends in SCDY aged 1—49 in Denmark over 20 years. Method All consecutive SCDY in Denmark from 2000-2019 were identified by a multi-source approach, where all deaths were screened using death certificates, autopsy reports, and nationwide registry data. Deaths in individuals aged 1-35 were assessed throughout the entire study period, whereas deaths aged 36-49 were assessed from 2007-2019. Patient characteristics and causes of death were compared throughout the study period. Annual incidence rates (IR) per 100.000 person-years (p-y) were calculated using Poisson regression and corrected for age and sex by direct standardization. Temporal changes in SCD incidences were computed as annual percentage changes (APC). Results During the study period, we identified 3.206 SCD aged 1—49 years; the mean age was 42 years [IQR 33-46], and 73% were male. Over the 20-year period, patient demographics and major causes of death remained similar. The rate of witnessed SCD declined significantly from 51% to 29% (p <0.001). For SCD aged 1—35, the standardized IR decreased from 2.98/100.000 p-y (95%CI 2.34;3.75) to 1.79/100.000 p-y (95%CI 1.30;2.40), equivalent to a 42% reduction and an APC of -3.6% (95%CI -2.7;-4.5). Among SCD aged 36-49, the IR declined from 24.57/100.000 p-y (95%CI 21.70;27.71) to 10.84/100.000 p-y (95%CI 8.94;13.03), equivalent to a 56% reduction and an APC of -5.9% (95%CI -3.9;-7.9); the intergroup difference was significant (p = 0.4). Conclusion The incidence rates of SCDY declined by almost 50% during the study period for all SCDY, but the decline was significantly higher among older individuals (36-49 years). Improved resuscitation of cardiac arrests has been a critical contributor to decreasing SCD mortality, as demonstrated by the significant decline in witnessed SCDY.Incidence rates of SCD in the young

A reproducible and replicable approach for harmonizing Landsat-8 and Sentinel-2 images

Clouds and cloud shadows significantly impact optical remote sensing. Combining images from different sources can help to obtain more frequent time series of the Earth’s surface. Nevertheless, sensor differences must be accounted for and treated before combining images from multiple sensors. Even after geometric correction, inter-calibration, and bandpass, disparities in image measurements can persist. One potential factor contributing to this phenomenon is directional effects. Bidirectional reflectance distribution function (BRDF) corrections have emerged as an optional processing method to soften differences in surface reflectance (SR) measurements, where the c-factor is one of the available options for this task. The c-factor efficiency is well-proven for medium spatial resolution products. However, its use should be restricted to images from sensors with a narrow view since it causes subtle changes in the processed images. There are currently a limited number of open tools for users to independently process their images. Here, we implemented the required tools to generate a Nadir BRDF-Adjusted Surface Reflectance (NBAR) product through the c-factor approach, and we evaluated them for a study area using Landsat-8 and Sentinel-2 images. Several comparisons were conducted to verify the SR and NBAR differences. Initially, a single-sensor approach was adopted and later a multi-source approach. Notably, NBAR products exhibit fewer disparities compared to SR products (prior to BRDF corrections). The results reinforce that the c-factor can be used to improve time series compatibility and, most importantly, provide the tools to allow users to generate the NBAR products themselves.

Differences among young unwitnessed sudden cardiac death, according to time from last seen alive: Insights from a 15-year nationwide study

More than half of all sudden cardiac deaths (SCD) are unwitnessed, but the composition of the unwitnessed SCD population is poorly described. This study aimed to compare clinical and autopsy characteristics of young unwitnessed SCD, depending on the time from last contact to being found dead. All unwitnessed SCD aged 1-35 years in Denmark from 2000-2014 identified through a multi-source approach were included. Time from last seen alive to being found dead was dichotomized in <1 hour or 1-24 hours. Clinical characteristics and autopsy results were compared, and predictors of autopsy were assessed by logistic regression. Of 440 unwitnessed SCD, 366 (83%) had not been seen alive within one hour of being found dead. Comorbidities differed between the groups with more epilepsy (17% vs 5%) and psychiatric diseases (13% vs 7%) in the 24-hour group. The 24-hour group died more frequently during sleep (64% vs 23%), the autopsy rate was higher (75% vs 61%), and deaths were more often unexplained after autopsy (69% vs 53%). Having been seen within one hour of death independently decreased the chance of being autopsied with OR 0.51 [95% CI 0.27-1.00, p = 0.0497]. The majority of unwitnessed SCD had not been seen alive within one hour of being found dead. Clinical- and autopsy-related characteristics differed among the two groups. Differences were mainly attributable to death-related circumstances and comorbidities. Excluding SCD cases not seen alive within one hour of being found dead would severely underestimate the burden of SCD.

THE DEVELOPMENT OF A 2D/3D BIM-GIS WEB PLATFORM FOR PLANNED MAINTENANCE OF BUILT AND CULTURAL HERITAGE: THE MAIN10ANCE PROJECT

Abstract. The integration of Building Information Modeling (BIM), or Historic BIM (HBIM), and Geographic Information Systems (GIS) has assumed a crucial role not only in the management of smart cities or urban districts, as well as in the construction sector or infrastructures but also in the management, conservation and safeguarding of cultural heritage. This integration opens up new opportunities for decision-makers to assess the impact of interventions on heritage sites, evaluate preservation strategies, and facilitate informed decision-making processes. This contribution presents and describes the open-source web platform developed within the Interreg MAIN10ANCE project. The project aims to promote a common methodology for the cooperation and collaboration of the various stakeholders involved in the restoration, management and maintenance activities of built and cultural heritage sites. This objective has been pursued by creating a BIM-GIS database, focused on the definition of planned maintenance activities, whose data has been made available to professionals and management bodies by means of a shared web platform, usable directly in situ on a tablet or portable devices. The paper highlights the key components of the web platform, including the BIM, GIS, and artifact Viewers, the Planner, and the overall Dashboard. The platform's multiscale and multi-source approach makes it extensible to various cultural assets, whether environmental, architectural, or movable. By integrating historical and geometric information and planning maintenance activities, the web platform thus supports public administrations and professionals, amending existing state-of-the-art solutions.

Assessment of ensemble learning for object-based land cover mapping using multi-temporal Sentinel-1/2 images

Accurate land cover mapping is challenging in Southeast Asia where cloud coverage is prevalent and landscape is heterogenous. Object-based mapping, multi-temporal images and combined use of optical and microwave data, provide abundant features in spectral, spatial, temporal, geometric and polarimetric dimensions. And random forest is usually employed due to robustness and efficiency in handling high-dimensional and noisy data. This study assesses whether feature selection and ensemble analysis, which are rarely adopted, yield improved result. Recursive feature elimination decreases original 568 features into a subset of 53 features, achieving the optimal combination of features. Ensemble analysis of random forest, support vector machine, and K-nearest neighbors leads to an overall accuracy of 0.816. Comparison experiments demonstrated the merits of the multi-temporal, multi-source approach, feature elimination and ensemble analysis.