Temporal Adaptive Reflectance Fusion Model Research Articles

Tracking crop phenology in a highly dynamic landscape with knowledge-based Landsat–MODIS data fusion

Earth observation image data are regularly used to capture surface conditions over large areas, but there is a trade-off between high (or low) spatial and low (or high) temporal resolution. The Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) overcomes this trade-off by fusing high spatial and temporal resolution multisource image data. However, ESTARFM requires additional modifications in order to provide reliable estimates of surface conditions showing large spectral differences in highly dynamic and fragmented agricultural systems. We modified ESTARFM by taking a knowledge-based approach to track maize and rice phenology in a highly dynamic and fragmented agricultural landscape in Ethiopia in 2019. The two major improvements included: (i) Selection of Landsat-MODIS imageries based on crop sowing and harvesting information and (ii) generation and use of a land cover map to select similar pixels. We assessed model performance with the enhanced vegetation index (EVI) derived from independent Landsat image data and in-situ leaf area index (LAI) data. The improved ESTARFM workflow resulted in reliable Landsat-MODIS prediction (R2 = 0.67, RMSE = 0.07) compared to the standard ESTARFM workflow (R2 = 0.54 RMSE = 0.01) during the rapid growth stage. Our modifications outperformed the standard implementation of ESTARFM according to LAI magnitude (R2 = 0.73–0.84 versus R2 = 0.58–0.64) and phenological timing (RMSE = 8 days verses RMSE = 12 days). Our modified application of ESTARFM serves as a basis for monitoring crop growth and development in highly dynamic and fragmented agricultural systems.

Spatial and Temporal Resolution Improvement of Actual Evapotranspiration Maps Using Landsat and MODIS Data Fusion

Producing daily actual evapotranspiration (ETa) maps with high spatial resolution has always been a challenge for remote sensing research. This study assessed the feasibility of producing daily ETa maps with a high spatial resolution (30 m) for the sugarcane farmlands of Amir Kabir Sugarcane Agro-industry (Khuzestan, Iran) using three different scenarios. In the first scenario, the reflectance bands of Landsat 8 were predicted from the moderate resolution imaging spectroradiometer (MODIS) imagery using the spatial and temporal adaptive reflectance fusion model (STARFM) algorithm. Also, the thermal bands of Landsat 8 were predicted by the spatiotemporal adaptive data fusion algorithm for temperature mapping (SADFAT). Then, ETa amounts were calculated employing such bands and the surface energy balance algorithm for land (SEBAL). In the second scenario, the input data needed by SEBAL were downscaled using the MODIS images and different methods. Then, using the downscaled data and SEBAL, daily ETa amounts with a spatial resolution of 30 m were calculated. In the third scenario, ETa data acquired by MODIS were downscaled to the scale of Landsat 8. In the second and third scenarios, downscaling of the data was carried out by the ratio, regression, and neural networks methods with two different approaches. In the first approach, the Landsat image on day 1 and the relationship between the two MODIS images on day 1 and the other days were used. In the second approach, the simulated image on the previous day and the relationship between the two consecutive images of MODIS were used. Comparing the simulated ETa amounts with the ETa amounts derived from Landsat 8, the first scenario had the best result with an RMSE (root mean square error) of 0.68 mm day−1. The neural networks method used in the third scenario with the second approach had the worst result with an RMSE of 2.25 mm day−1, which was however a better result than the ETa amounts derived from MODIS with an RMSE of 3.19 mm day−1. The method developed in this study offers an efficient and inexpensive way to produce daily ETa maps with a high spatial resolution. Furthermore, we suggest that STARFM and SADFAT algorithms have acceptable accuracies in the simulation of reflectance and thermal bands of Landsat 8 images for homogeneous areas.

Evaluation of Landsat 8-like Land Surface Temperature by Fusing Landsat 8 and MODIS Land Surface Temperature Product

High-spatiotemporal-resolution land surface temperature (LST) is a crucial parameter in various environmental monitoring. However, due to the limitation of sensor trade-off between the spatial and temporal resolutions, such data are still unavailable. Therefore, the generation and verification of such data are of great value. The spatiotemporal fusion algorithm, which can be used to improve the spatiotemporal resolution, is widely used in Landsat and MODIS data to generate Landsat-like images, but there is less exploration of combining long-time series MODIS LST and Landsat 8 LST product to generate Landsat 8-like LST. The purpose of this study is to evaluate the accuracy of the long-time series Landsat 8 LST product and the Landsat 8-like LST generated by spatiotemporal fusion. In this study, based on the Landsat 8 LST product and MODIS LST product, Landsat 8-like LST is generated using Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), Enhanced STARFM (ESTARFM), and the Flexible Spatiotemporal DAta Fusion (FSDAF) algorithm, and tested and verified in the research area located in Gansu Province, China. In this process, Landsat 8 LST product was verified based on ground measurements, and the fusion results were comprehensively evaluated based on ground measurements and actual Landsat 8 LST images. Ground measurements verification indicated that Landsat 8 LST product was highly consistent with ground measurements. The Root Mean Square Error (RMSE) was 2.862 K, and the coefficient of determination R2 was 0.952 at All stations. Good fusion results can be obtained for the three spatiotemporal algorithms, and the ground measurements verified at All stations show that R2 was more significant than 0.911. ESTARFM had the best fusion result (R2 = 0.915, RMSE = 3.661 K), which was better than STARFM (R2 = 0.911, RMSE = 3.746 K) and FSDAF (R2 = 0.912, RMSE = 3.786 K). Based on the actual Landsat 8 LST images verification, the fusion images were highly consistent with actual Landsat 8 LST images. The average RMSE of fusion images about STARFM, ESTARFM, and FSDAF were 2.608 K, 2.245 K, and 2.565 K, respectively, and ESTARFM is better than STARFM and FSDAF in most cases. Combining the above verification, the fusion results of the three algorithms were reliable and ESTARFM had the highest fusion accuracy.

Extraction of crop information through the spatiotemporal fusion of OLI and MODIS images

Spatiotemporal data fusion algorithms have been developed to fuse satellite imagery from sensors with different spatial and temporal resolutions and generate predicted imagery. In this study, we compare the predictions of three spatiotemporal data fusion algorithms in blending Landsat-8/OLI and Terra-Aqua/MODIS images for mapping soybean and corn under five classification scenarios. The Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM), and Flexible Spatiotemporal Data Fusion (FSDAF) algorithms were compared to generate images for the 2016/2017 summer crop-year. Classifications including phenological metrics extracted from FSDAF- and STARFM-predicted EVI time series had overalls accuracies higher than the other scenarios, 93.11% and 91.33%, respectively. The results show that phenological metrics extracted from predicted images are an interesting alternative to overcome cloud cover frequency limitations for soybean and corn mapping in tropical areas.

Forest classification using synthetic GF-1/WFV time series and phenological parameters

High-accuracy forest type mapping is crucial for forest resource inventory to support forest management, conservation biology, and ecological restoration. The Chinese GF-1 satellite can provide high-spatial resolution and low-cost data for forestry monitoring. However, it is difficult to obtain time-series GF-1 images with high spatial resolution and high temporal frequency because of technical constraints and cloud contamination, which affects the accuracy of forest type extraction. We proposed a method for mapping forest type using Chinese GF-1/widefield view (WFV) multi-spectral data, the fused GF-1/WFV normalized difference vegetation index (NDVI) time series, and the phenological parameters. We used the enhanced spatial and temporal adaptive reflectance fusion model algorithm to fuse Chinese GF-1/WFV and Moderate Resolution Imaging Spectroradiometer NDVI to generate high-spatiotemporal resolution NDVI time-series image data and extract phenological parameters in the northeastern part of Hubei Province, China. We then designed four different combinations of GF-1/WFV spectral features, the fused NDVI time-series features, and the phenological parameters. Finally, a random forest algorithm is used to classify forest types in the study area. The results showed that the proposed method can obtain satisfactory results. The overall accuracy based on the combination of GF-1/WFV spectral features, high-resolution time-series NDVI, and phenological features was 89.68%. Compared with the classification using only GF-1/WFV spectral data, the overall accuracy was improved by 9.60% points. These results demonstrate that the proposed method can support forest type survey and mapping at the regional or global scale.

A boundary finding-based spatiotemporal fusion model for vegetation index

ABSTRACT The spatiotemporal image fusion technique is regarded as an effective and efficient remedy to blend two or more geometrically registered remote sensing images of the same scene acquired from different sensors into a single image with both high temporal and high spatial resolutions. A number of existing fusion methods developed for generating Landsat-like images from the Moderate Resolution Imaging Spectroradiometer (MODIS) have successfully solved limitations on vegetation dynamics monitoring. However, these methods either treat the weighting parameters as constants or adopt insufficient linear regression coefficients (considering only spatial, temporal or spectral weights), which may contribute to blocky artefacts and dissimilar pixels in the predicted image. In this study, a boundary finding-based vegetation index spatiotemporal fusion model (BESFM) is newly proposed to generate high spatial and high temporal resolutions Enhanced Vegetation Index (EVI) images in homogeneous and heterogeneous regions with rapid vegetation changes. The cloud-free Landsat-8 Operational Land Imager (OLI) and the MODIS (MOD09GA) data acquired from two study sites, located in Queensland and New South Wales and Australia, were selected to evaluate the performance of the proposed method. Compared with the enhanced spatial and temporal adaptive reflectance fusion model (ESTARFM) and flexible spatiotemporal data fusion (FSDAF), the BESFM outperforms in both visible image quality (reducing the spatial distortion and blocky artefacts in prediction caused by abrupt land cover changes) and quantitative indices (with RSME minimum to 0.011 and R2 maximum to 0.956). Since the BESFM method not only uses adjustable linear regression assumptions for heterogeneous and homogeneous regions differently but also implements boundary detection-based spatial weight with temporal and spectral weights together to enhance the predicted spatial details, it has the potential to increase the availability of fine spatial resolution data.

Feasibility of the Spatiotemporal Fusion Model in Monitoring Ebinur Lake’s Suspended Particulate Matter under the Missing-Data Scenario

High-frequency monitoring of suspended particulate matter (SPM) concentration can improve water resource management. Missing high-resolution satellite images could hamper remote-sensing SPM monitoring. This study resolved the problem by applying spatiotemporal fusion technology to obtain high spatial resolution and dense time-series data to fill image-data gaps. Three data sources (MODIS, Landsat 8, and Sentinel 2) and two spatiotemporal fusion methods (the enhanced spatial and temporal adaptive reflectance fusion model (ESTARFM) and the flexible spatiotemporal data fusion (FSDAF)) were used to reconstruct missing satellite images. We compared their fusion accuracy and verified the consistency of fusion images between data sources. For the fusion images, we used random forest (RF) and XGBoost as inversion methods and set “fusion first” and “inversion first” strategies to test the method’s feasibility in Ebinur Lake, Xinjiang, arid northwestern China. Our results showed that (1) the blue, green, red, and NIR bands of ESTARFM fusion image were better than FSDAF, with a good consistency (R2 ≥ 0.54) between the fused Landsat 8, Sentinel 2 images, and their original images; (2) the original image and fusion image offered RF inversion effect better than XGBoost. The inversion accuracy based on Landsat 8 and Sentinel 2 were R2 0.67 and 0.73, respectively. The correlation of SPM distribution maps of the two data sources attained a good consistency of R2 0.51; (3) in retrieving SPM from fused images, the “fusion first” strategy had better accuracy. The optimal combination was ESTARFM (Landsat 8)_RF and ESTARFM (Sentinel 2)_RF, consistent with original SPM maps (R2 = 0.38, 0.41, respectively). Overall, the spatiotemporal fusion model provided effective SPM monitoring under the image-absence scenario, with good consistency in the inversion of SPM. The findings provided the research basis for long-term and high-frequency remote-sensing SPM monitoring and high-precision smart water resource management.

ERTFM: An Effective Model to Fuse Chinese GF-1 and MODIS Reflectance Data for Terrestrial Latent Heat Flux Estimation

Coarse spatial resolution sensors play a major role in capturing temporal variation, as satellite images that capture fine spatial scales have a relatively long revisit cycle. The trade-off between the revisit cycle and spatial resolution hinders the access of terrestrial latent heat flux (LE) data with both fine spatial and temporal resolution. In this paper, we firstly investigated the capability of an Extremely Randomized Trees Fusion Model (ERTFM) to reconstruct high spatiotemporal resolution reflectance data from a fusion of the Chinese GaoFen-1 (GF-1) and the Moderate Resolution Imaging Spectroradiometer (MODIS) products. Then, based on the merged reflectance data, we used a Modified-Satellite Priestley–Taylor (MS–PT) algorithm to generate LE products at high spatial and temporal resolutions. Our results illustrated that the ERTFM-based reflectance estimates showed close similarity with observed GF-1 images and the predicted NDVI agreed well with observed NDVI at two corresponding dates (r = 0.76 and 0.86, respectively). In comparison with other four fusion methods, including the widely used spatial and temporal adaptive reflectance fusion model (STARFM) and the enhanced STARFM, ERTFM had the best performance in terms of predicting reflectance (SSIM = 0.91; r = 0.77). Further analysis revealed that LE estimates using ERTFM-based data presented more detailed spatiotemporal characteristics and provided close agreement with site-level LE observations, with an R2 of 0.81 and an RMSE of 19.18 W/m2. Our findings suggest that the ERTFM can be used to improve LE estimation with high frequency and high spatial resolution, meaning that it has great potential to support agricultural monitoring and irrigation management.

Evaluating a spatiotemporal shape-matching model for the generation of synthetic high spatiotemporal resolution time series of multiple satellite data

Time series of high spatiotemporal resolution satellite data are required to monitor land surface biophysical properties and their seasonal and inter-annual dynamics at sub-field to field scales. To generate such time series, various algorithms have been developed to fuse infrequent cloud-free Landsat observations with daily Moderate Resolution Imaging Spectroradiometer (MODIS) observations. An early and widely used Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) was developed based on the assumptions that observations from the two sensors are consistent in terms of spatial aggregation and temporal variation. These assumptions are not always valid in the real world, especially in a complex heterogeneous region. In this study, we investigated a Spatiotemporal Shape-Matching Model (SSMM) to generate synthetic time series of high spatiotemporal resolution satellite data. The SSMM, which is conceptually different from the image pair-based STARFM and STARFM-like approaches, makes full use of all spatiotemporally matched fine and coarse resolution data in an entire time series to establish a temporally uniformed fusion model for a given fine resolution pixel. The SSMM assumes that the temporal shapes of both the fine and coarse resolution time series are similar, but their magnitudes and phenological phases could differ largely even for the same vegetation type. This study assessed the capability of the SSMM to generate high spatiotemporal resolution time series of two-band Enhanced Vegetation Index (EVI2). Specifically, we generated the synthetic 30 m time series using the SSMM and STARFM algorithms in the northeastern United States based on 500 m Visible Infrared Imaging Radiometer Suite (VIIRS) data and 30 m Landsat 8 and Sentinel-2 observations. We then evaluated the SSMM-derived 30 m time series across 15 land cover types and various degrees of heterogeneity. The result indicates that the SSMM is able to effectively generate synthetic time series in all different land cover types, which has advantages over the STARFM approach. Although the SSMM performance is degraded in heterogeneous regions, it can explain 82–91% of variations in 30 m EVI2 time series and produce constant root mean square error (0.053–0.056) across various levels of heterogeneity. Moreover, the SSMM can explain 87%-93% and 69%-87% of variation in 30 m EVI2 time series, respectively, with the models established using more than 10 pairs and 4–10 pairs of fine and coarse resolution observations. This suggests that the SSMM is also capable of generating high spatiotemporal resolution time series using historical Landsat time series with limited cloud-free observations, which are critical for studying vegetation dynamics and monitoring crop conditions.

Precipitation events determine the spatiotemporal distribution of playa surface salinity in arid regions: evidence from satellite data fused via the enhanced spatial and temporal adaptive reflectance fusion model

Playas are desert landscapes unique to arid regions that respond quickly to climate change. Changes in surface salinity in Praia are directly linked to water security and regional ecological and economic health. This study aimed to explore the sensitivity of remotely sensed spectral indices to precipitation events. An ESTARFM was utilized to fuse daily (less than 10% cloudiness) MODIS and Landsat 8 imagery from April to September 2019 in the Ebinur Lake Wetland Reserve. Salinity was tested in the laboratory after soil collections in May (spring) and August (summer). The significantly correlated normalized difference vegetation and salinity indices were compared to the precipitation data and found to fluctuate in response to precipitation. The normalized difference vegetation and salinity indices decreased following precipitation events and increased during precipitation intervals, and the precipitation amount and evaporation intensities may have influenced the magnitudes of the decrease and increase. Following precipitation, the playa surface produces new puddles and a new spatial distribution pattern of soil salinity. This study determined that the spatiotemporal fusion technique is an effective method for observing the dynamics of playa surfaces, and precipitation and evaporation affect the spatial distribution of salt on the playa surface; however, the monitoring period should be short when utilizing remote sensing to monitor playa salinity.

Improving the spatiotemporal fusion accuracy of fractional vegetation cover in agricultural regions by combining vegetation growth models

Spatiotemporal fusion has provided a feasible way to generate fractional vegetation cover (FVC) data with high spatial and temporal resolution. However, when the currently available spatiotemporal fusion methods are applied over agricultural regions, they usually underestimate high FVC values at the peak vegetation growth stage with medium FVC values as base data. This mainly results from inconsistencies in the temporal variations between fine- and coarse-resolution data if substantial temporal changes occur in vegetation. Therefore, a Spatial and Temporal Fusion method combining with Vegetation Growth Models (STF-VGM) was proposed to address this problem in this study, which incorporates vegetation growth models into the fusion process. Unlike other spatiotemporal fusion methods that mainly rely on changes in coarse-resolution data for prediction, STF-VGM fully utilizes available coarse- and fine-resolution time series data, including uncontaminated information in cloud/cloud shadow contaminated images. By establishing vegetation growth models with time series data, a conversion relationship between coarse- and fine-resolution FVC that changes along with the nonlinear vegetation change process can be extracted. STF-VGM makes prediction based on this variable relationship. A typical agricultural region located in the North China Plain was selected as the study area. The validation results indicated that the prediction accuracy for high FVC values was significantly improved using STF-VGM compared to the commonly used Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) and Flexible Spatiotemporal DAta Fusion (FSDAF) methods (STF-VGM: coefficient of determination (R2) = 0.9491, root mean square error (RMSE) = 0.0650, average difference (AD) = -0.0092; ESTARFM: R2 = 0.9341, RMSE = 0.1127, AD = -0.0631; FSDAF: R2 = 0.9224, RMSE = 0.1110, AD = -0.0599). The satisfactory performance of STF-VGM was also achieved in predicting FVC values at other vegetation growth stages (early growth stage: R2 = 0.8277, RMSE = 0.0440, AD = 0.0027; rapid growth stage: R2 = 0.9183, RMSE = 0.0844, AD = 0.0500). In addition, STF-VGM also has the potential to improve the spatiotemporal fusion accuracy of other vegetation parameters and vegetation indices, which will be evaluated in the future.

Determining the impacts of climate change and urban expansion on net primary productivity using the spatio-temporal fusion of remote sensing data

Climate change (CLC) and urban expansion (URE) have profoundly altered the terrestrial net primary productivity (NPP). Many studies have determined the effects of CLC and URE on the NPP. However, these studies were conducted at low resolutions (250–1000 m), making it difficult to detect many smaller new urban lands, and thus potentially underestimating the contribution of URE. To accurately determine the contributions of CLC and URE to the NPP, this study takes Beijing as an example and uses an Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) to fuse the spatial resolution of the Landsat Normalized Difference Vegetation Index (NDVI) and the temporal resolution of the Moderate Resolution Imaging Spectroradiometer (MODIS) NDVI to generate a new NDVI with a high spatio-temporal resolution. Compared with the Landsat NDVI, the NDVI fused by the ESTARFM is found to be reliable. The fused NDVI was then inputted into the Carnegie–Ames–Stanford Approach (CASA) model to generate the NPP with a high spatio-temporal resolution, namely, the 30-m NPP. Compared with the 250-m NPP generated by directly inputting the MODIS NDVI into the CASA model, the 30-m NPP as a new ecological indicator is more accurate than the 250-m NPP. Due to the high resolution of the 30-m NPP and its increased ability to detect more new urban lands, the total loss of the 30-m NPP caused by URE is much higher than that of the 250-m NPP. For the same reason, especially in rapidly urbanized areas, the contribution ratio of URE to the 30-m NPP is much higher than that to the 250-m NPP. Moreover, in natural vegetation cover areas, CLC, which is measured by the interannual changes in temperature, precipitation, and solar radiation, is the leading factor of the change in the NPP. However, within the urban areas, residual factors other than CLC and URE, such as the introduction of exotic high-productivity vegetation, irrigation, fertilization, and pest control, dominate the change in the NPP. The results of this study are expected to contribute to a deeper understanding of the influences of CLC and URE on terrestrial ecosystem carbon cycles and provide an important theoretical reference for urban planning.

Fusing Precipitable Water Vapor Data in CHINA at Different Timescales Using an Artificial Neural Network

Global climate change has noticeable influences on the water vapor redistribution in China, which is embodied by the fact that both wetting and drying tendencies were observed across China. This poses the necessity to monitor and understand the water vapor evolution in China. However, observations of water vapor from different techniques are subjected to systematic biases, different spatiotemporal resolutions and coverages, and different accuracy, which would hamper their joint use, potentially leading to contradictory conclusions when using different techniques. Data fusion is a promising way to address this problem. Some scholars have proposed several methods to fuse multi-source PWV data in China region, such as the enhanced spatial and temporal adaptive reflectance fusion model, the hybrid PWV fusion model, and the linear calibration model. Although these models can produce PWV products with improved accuracy, they still have some shortcomings, such as no consideration for spatial or temporal variations in bias or inevitably impose some biases inaccurate information since assumptions made for interpolations are imperfect. In this study, we use the high-quality Global Navigation Satellite System (GNSS) precipitable water vapor (PWV) to calibrate and optimize the Moderate-resolution Imaging Spectroradiometer (MODIS) and the European Centre for Medium-Range Weather Forecasts ReAnalyses 5 (ERA5) PWV in 2018–2019 through a Generalized Regression Neural Network (GRNN) at annual, quarterly, and monthly timescales. Validation results demonstrate that modifying the MODIS and ERA5 PWV at the monthly timescale results in the best accuracy. In the monthly experiment, the average bias, standard deviation (STD), and root mean square (RMS) error of modified MODIS PWV are 0.0 mm, 2.6 mm, and 2.6 mm, respectively. The percentage improvement is as high as 50% in terms of RMS compared to the original MODIS PWV. It becomes 0.0 mm, 1.7 mm, and 1.7 mm for the modified ERA5 PWV and the percentage improvement is 40%. Since the biases among different products are well-calibrated and the accuracy of MODIS and ERA5 PWV is improved to the same level of GNSS PWV, we can fuse them by simply merging them. Finally, we generate a new product of PWV in China with a temporal resolution of 1 day, a spatial resolution better than 31 km, and an accuracy better than 2.7 mm, which will serve as a high-quality product for investigating the water vapor redistribution under a changing climate.

On-the-Fly Fusion of Remotely-Sensed Big Data Using an Elastic Computing Paradigm with a Containerized Spark Engine on Kubernetes.

Remotely-sensed satellite image fusion is indispensable for the generation of long-term gap-free Earth observation data. While cloud computing (CC) provides the big picture for RS big data (RSBD), the fundamental question of the efficient fusion of RSBD on CC platforms has not yet been settled. To this end, we propose a lightweight cloud-native framework for the elastic processing of RSBD in this study. With the scaling mechanisms provided by both the Infrastructure as a Service (IaaS) and Platform as a Services (PaaS) of CC, the Spark-on-Kubernetes operator model running in the framework can enhance the efficiency of Spark-based algorithms without considering bottlenecks such as task latency caused by an unbalanced workload, and can ease the burden to tune the performance parameters for their parallel algorithms. Internally, we propose a task scheduling mechanism (TSM) to dynamically change the Spark executor pods’ affinities to the computing hosts. The TSM learns the workload of a computing host. Learning from the ratio between the number of completed and failed tasks on a computing host, the TSM dispatches Spark executor pods to newer and less-overwhelmed computing hosts. In order to illustrate the advantage, we implement a parallel enhanced spatial and temporal adaptive reflectance fusion model (PESTARFM) to enable the efficient fusion of big RS images with a Spark aggregation function. We construct an OpenStack cloud computing environment to test the usability of the framework. According to the experiments, TSM can improve the performance of the PESTARFM using only PaaS scaling to about 11.7%. When using both the IaaS and PaaS scaling, the maximum performance gain with the TSM can be even greater than 13.6%. The fusion of such big Sentinel and PlanetScope images requires less than 4 min in the experimental environment.

A practical reanalysis data and thermal infrared remote sensing data merging (RTM) method for reconstruction of a 1-km all-weather land surface temperature

An all-weather land surface temperature (LST) dataset at moderate to high spatial resolutions (e.g. 1 km) has been in urgent need, especially in areas frequently covered in clouds (i.e. the Tibetan Plateau). Merging satellite thermal infrared (TIR) and passive microwave (PMW) observations is a widely-adopted approach to derive such LST datasets, whereas the swath gap of the PMW data leads to considerable data deficiency or low reliability in the merged LST, especially at the low-mid latitudes. Fortunately, reanalyzed data provides the spatiotemporally continuous LST and thus, is promising to be merged with TIR data for reconstructing the all-weather LST without this issue. However, few studies along this direction have been reported. In this context, based on the decomposition model of LST time series, this study proposes a novel reanalysis and thermal infrared remote sensing data merging (RTM) method to reconstruct the 1-km all-weather LST. The method was applied to merge Aqua/Terra Moderate Resolution Imaging Spectroradiometer (MODIS) and Global/China Land Data Assimilation System (GLDAS/CLDAS) data over the Tibetan Plateau and the surrounding area. Results show that the RTM LST has RMSEs of 2.03–3.98 K and coefficients of determination of 0.82–0.93 under all-weather conditions when validated against the ground measured LST. Besides, from comparison between RTM LST and current existing PMW-TIR merged LST, it is found the former LST efficiently outperforms the latter one in terms of accuracy and image quality, especially over the MW swath gap-covered area. In addition, compared to the MODIS-CLDAS merged all-weather LST based on the Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM), the two LSTs have comparable accuracy while the RTM LST has higher spatial completeness. The method is promising for generating a long-term all-weather LST record at moderate to high spatiotemporal resolutions at large scales, which would be beneficial to associated studies and applications.

Spatiotemporal fusion method to simultaneously generate full-length normalized difference vegetation index time series (SSFIT)

High spatiotemporal resolution normalized difference vegetation index (NDVI) time-series imagery is required for monitoring vegetation dynamics with dense observations and spatial details, and in recent years, many spatiotemporal data fusion methods have been proposed to fulfill this need. However, strict data requirements and inappropriate modeling strategies often limit their performance, particularly under poor conditions of available input data. In this study, we proposed a Spatiotemporal fusion method to Simultaneously generate Full-length normalized difference vegetation Index Time series (SSFIT) with a high spatial resolution and frequent coverage. The utilization of temporal information across sensors contributes to its two distinct features: (1) no cloud-free high-spatial-resolution image is required and (2) high-spatial-resolution images on multiple prediction dates are generated at the same time. Comparison experiments were conducted by simulating ideal and challenging conditions of input time-series data in two characteristic areas, and the proposed methods were also compared with four typical methods: The Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), Flexible Spatiotemporal DAta Fusion (FSDAF), Fit-FC (regression model Fitting, spatial Filtering and residual Compensation), and Improved FSDAF. The results demonstrate that SSFIT yields a better overall prediction accuracy and efficiency under ideal input conditions (average root mean square error of 0.1037 and 0.0713, average correlation coefficient of 0.9180 and 0.7875, and computation time of 109 and 120 s, computed in the two test areas) and is more robust against the decrease in available input data under challenging conditions. SSFIT is thus expected to be extended to various remote sensing products and support applications for monitoring land surface dynamics.

Estimation of Grassland Carrying Capacity by Applying High Spatiotemporal Remote Sensing Techniques in Zhenglan Banner, Inner Mongolia, China

Overgrazing directly leads to grassland degradation, which is a serious constraint to the sustainable development of animal husbandry. In drylands, grassland biomass is highly heterogeneous in space and time. It is difficult to achieve sustainable utilization of grassland resources by focusing only on the average annual carrying capacity assessment obtained from grass yield. Here, we proposed a novel approach for assessing grassland carrying capacity, taking Zhenglan Banner (County) in Inner Mongolia as the study area. First, monthly grass yield at 30 m spatial resolution was estimated, derived from Carnegie–Ames–Stanford Approach (CASA) model and spatial and temporal adaptive reflectance fusion model (STARFM). Then, based on the degree of sand mobility and degradation condition of typical steppe, the utilization patterns for sandy land and typical steppe in different grazing seasons were developed separately to obtain available grass yield. Finally, the carrying capacity at the Gacha (Village)-scale was estimated and the current livestock carrying status was evaluated to facilitate the grassland refined management. In Zhenglan Banner, the carrying capacity was 237.46 thousand cattle-units in summer. The grassland resources are being overgrazed, with an overload rate of 19.32%. At Gacha-scale, the maximum reasonable stock density was ranged from 0.06 cattle-unit/ha to 0.42 cattle-unit/ha. Fifty-one Gachas exhibited livestock overload. This study is expected to provide technical support and scientific reference data for ecological conservation and grassland management in the study area, as well as in dryland pastoral areas of northern China.

Phenological Analysis of Sub-Alpine Forest on Jeju Island, South Korea, Using Data Fusion of Landsat and MODIS Products

Climate change poses a disproportionate risk to alpine ecosystems. Effective monitoring of forest phenological responses to climate change is critical for predicting and managing threats to alpine populations. Remote sensing can be used to monitor forest communities in dynamic landscapes for responses to climate change at the species level. Spatiotemporal fusion technology using remote sensing images is an effective way of detecting gradual phenological changes over time and seasonal responses to climate change. The spatial and temporal adaptive reflectance fusion model (STARFM) is a widely used data fusion algorithm for Landsat and MODIS imagery. This study aims to identify forest phenological characteristics and changes at the species–community level by fusing spatiotemporal data from Landsat and MODIS imagery. We fused 18 images from March to November for 2000, 2010, and 2019. (The resulting STARFM-fused images exhibited accuracies of RMSE = 0.0402 and R2 = 0.795. We found that the normalized difference vegetation index (NDVI) value increased with time, which suggests that increasing temperature due to climate change has affected the start of the growth season in the study region. From this study, we found that increasing temperature affects the phenology of these regions, and forest management strategies like monitoring phenology using remote sensing technique should evaluate the effects of climate change.

Fusion of GF and MODIS Data for Regional-Scale Grassland Community Classification with EVI2 Time-Series and Phenological Features

Satellite-borne multispectral data are suitable for regional-scale grassland community classification owing to comprehensive coverage. However, the spectral similarity of different communities makes it challenging to distinguish them based on a single multispectral data. To address this issue, we proposed a support vector machine (SVM)–based method integrating multispectral data, two-band enhanced vegetation index (EVI2) time-series, and phenological features extracted from Chinese GaoFen (GF)-1/6 satellite with (16 m) spatial and (2 d) temporal resolution. To obtain cloud-free images, the Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) algorithm was employed in this study. By using the algorithm on the coarse cloudless images at the same or similar time as the fine images with cloud cover, the cloudless fine images were obtained, and the cloudless EVI2 time-series and phenological features were generated. The developed method was applied to identify grassland communities in Ordos, China. The results show that the Caragana pumila Pojark, Caragana davazamcii Sanchir and Salix schwerinii E. L. Wolf grassland, the Potaninia mongolica Maxim, Ammopiptanthus mongolicus S. H. Cheng and Tetraena mongolica Maxim grassland, the Caryopteris mongholica Bunge and Artemisia ordosica Krasch grassland, the Calligonum mongolicum Turcz grassland, and the Stipa breviflora Griseb and Stipa bungeana Trin grassland are distinguished with an overall accuracy of 87.25%. The results highlight that, compared to multispectral data only, the addition of EVI2 time-series and phenological features improves the classification accuracy by 9.63% and 14.7%, respectively, and even by 27.36% when these two features are combined together, and indicate the advantage of the fine images in this study, compared to 500 m moderate-resolution imaging spectroradiometer (MODIS) data, which are commonly used for grassland classification at regional scale, while using 16 m GF data suggests a 23.96% increase in classification accuracy with the same extracted features. This study indicates that the proposed method is suitable for regional-scale grassland community classification.

A Hybrid Deep Learning-Based Spatiotemporal Fusion Method for Combining Satellite Images with Different Resolutions

Spatiotemporal fusion (STF) is considered a feasible and cost-effective way to deal with the trade-off between the spatial and temporal resolution of satellite sensors, and to generate satellite images with high spatial and high temporal resolutions. This is achieved by fusing two types of satellite images, i.e., images with fine temporal but rough spatial resolution, and images with fine spatial but rough temporal resolution. Numerous STF methods have been proposed, however, it is still a challenge to predict both abrupt landcover change, and phenological change, accurately. Meanwhile, robustness to radiation differences between multi-source satellite images is crucial for the effective application of STF methods. Aiming to solve the abovementioned problems, in this paper we propose a hybrid deep learning-based STF method (HDLSFM). The method formulates a hybrid framework for robust fusion with phenological and landcover change information with minimal input requirements, and in which a nonlinear deep learning-based relative radiometric normalization, a deep learning-based superresolution, and a linear-based fusion are combined to address radiation differences between different types of satellite images, landcover, and phenological change prediction. Four comparative experiments using three popular STF methods, i.e., spatial and temporal adaptive reflectance fusion model (STARFM), flexible spatiotemporal data fusion (FSDAF), and Fit-FC, as benchmarks demonstrated the effectiveness of the HDLSFM in predicting phenological and landcover change. Meanwhile, HDLSFM is robust for radiation differences between different types of satellite images and the time interval between the prediction and base dates, which ensures its effectiveness in the generation of fused time-series data.