On the blending of the Landsat and MODIS surface reflectance: predicting daily Landsat surface reflectance. IEEE Trans Geosci Remote Sens

Earth Resources Technology Inc, Jessup, MD
IEEE Transactions on Geoscience and Remote Sensing (Impact Factor: 3.51). 08/2006; 44(8):2207-2218. DOI: 10.1109/TGRS.2006.872081
Source: DBLP


The 16-day revisit cycle of Landsat has long limited its use for studying global biophysical processes, which evolve rapidly during the growing season. In cloudy areas of the Earth, the problem is compounded, and researchers are fortunate to get two to three clear images per year. At the same time, the coarse resolution of sensors such as the Advanced Very High Resolution Radiometer and Moderate Resolution Imaging Spectroradiometer (MODIS) limits the sensors' ability to quantify biophysical processes in heterogeneous landscapes. In this paper, the authors present a new spatial and temporal adaptive reflectance fusion model (STARFM) algorithm to blend Landsat and MODIS surface reflectance. Using this approach, high-frequency temporal information from MODIS and high-resolution spatial information from Landsat can be blended for applications that require high resolution in both time and space. The MODIS daily 500-m surface reflectance and the 16-day repeat cycle Landsat Enhanced Thematic Mapper Plus (ETM+) 30-m surface reflectance are used to produce a synthetic "daily" surface reflectance product at ETM+ spatial resolution. The authors present results both with simulated (model) data and actual Landsat/MODIS acquisitions. In general, the STARFM accurately predicts surface reflectance at an effective resolution close to that of the ETM+. However, the performance depends on the characteristic patch size of the landscape and degrades somewhat when used on extremely heterogeneous fine-grained landscapes

Download full-text


Available from: Forrest G. Hall, Jan 31, 2015
    • "For the latter, the calculated thresholds are affected by the parameterization of the number of spectral slices (s), which provides an indication of the number of land cover classes in the image. The use of a higher s value implies a stricter criterion for selecting pixels within the local moving window that are similar in magnitude to the central pixel (Gao et al., 2006). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Satellite remote sensing has been used successfully to map leaf area index (LAI) across landscapes, but advances are still needed to exploit multi-scale data streams for producing LAI at both high spatial and temporal resolution. A multi-scale Spatio-Temporal Enhancement Method for medium resolution LAI (STEM-LAI) has been developed to generate 4-day time-series of Landsat-scale LAI from existing medium resolution LAI products. STEM-LAI has been designed to meet the demands of applications requiring frequent and spatially explicit information, such as effectively resolving rapidly evolving vegetation dynamics at sub-field (30 m) scales. In this study, STEM-LAI is applied to Moderate Resolution Imaging Spectroradiometer (MODIS) based LAI data and utilizes a reference-based regression tree approach for producing MODIS-consistent, but Landsat-based, LAI. The Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) is used to interpolate the downscaled LAI between Landsat acquisition dates, providing a high spatial and temporal resolution improvement over existing LAI products. STARFM predicts high resolution LAI by blending MODIS and Landsat based information from a common acquisition date, with MODIS data from a prediction date. To demonstrate its capacity to reproduce fine-scale spatial features observed in actual Landsat LAI, the STEM-LAI approach is tested over an agricultural region in Nebraska. The implementation of a 250 m resolution LAI product, derived from MODIS 1 km data and using a scale consistent approach based on the Normalized Difference Vegetation Index (NDVI), is found to significantly improve accuracies of spatial pattern prediction, with the coefficient of efficiency (E) ranging from 0.77–0.94 compared to 0.01–0.85 when using 1 km LAI inputs alone. Comparisons against an 11-year record of in-situ measured LAI over maize and soybean highlight the utility of STEM-LAI in reproducing observed LAI dynamics (both characterized by r2 = 0.86) over a range of plant development stages. Overall, STEM-LAI represents an effective downscaling and temporal enhancement mechanism that predicts in-situ measured LAI better than estimates derived through linear interpolation between Landsat acquisitions. This is particularly true when the in-situ measurement date is greater than 10 days from the nearest Landsat acquisition, with prediction errors reduced by up to 50%. With a streamlined and completely automated processing interface, STEM-LAI represents a flexible tool for LAI disaggregation in space and time that is adaptable to different land cover types, landscape heterogeneities, and cloud cover conditions.
    No preview · Article · May 2016 · International Journal of Applied Earth Observation and Geoinformation
  • Source
    • "STARFM relies on the assumption that land cover does not change between the estimation-and reference-time periods. A series of weights (spatial, temporal, and distance weights) were introduced to increase the ability of the fusion model to detect land cover change (Gao et al., 2006). STARFM has been proven useful to detect gradual changes but was shown to be less effective in detecting abrupt changes often caused by disturbances (Hilker et al., 2009). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Monitoring forest disturbances using remote sensing data with high spatial and temporal resolution can reveal relationships between forest disturbances and forest ecological patterns and processes. In this study, we fused Landsat data at high spatial resolution (30 m) with 8-day MODIS data to produce high spatial and temporal resolution image time-series. The Spatial Temporal Adaptive Algorithm for mapping Reflectance Change (STAARCH) is a simple but effective fusion method. We adapted the STAARCH fusion method to successfully produce a time-series of disturbances with high overall accuracy (89–92%) in mixed forests in southeast Oklahoma. The results demonstrated that in southeast Oklahoma, forest area disturbed in 2011 was higher than it was in 2000. However, two remarkable drops were identified in 2001 and 2006. We speculated that the drops were related to the economic recessions causing reduction in the demand of woody products. The detected fluctuation of area disturbed calls for continuing monitoring of spatial and temporal changes in this and other forest landscapes using high spatial and temporal resolution imagery datasets to better recognize the economic and environmental factors, as well as the consequences of those changes. (Free access until 9/23 - http://authors.elsevier.com/a/1RU9j14ynS3AVi)
    Full-text · Article · Feb 2016 · International Journal of Applied Earth Observation and Geoinformation
  • Source
    • "However, they face challenges in heterogeneous regions with abrupt land cover type changes. Most weighted function based methods assume no land cover type changes between input and prediction date (Fu et al., 2013; Gao et al., 2006; Weng, Fu, & Gao, 2014; Zhu et al., 2010). As a result, they can successfully predict pixels with changes in attributes like vegetation phenology or soil moisture, because these changes are strongly related to the changes in similar pixels selected from the input imagery. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Studies of land surface dynamics in heterogeneous landscapes often require remote sensing data with high acquisition frequency and high spatial resolution. However, no single sensor meets this requirement. This study presents a new spatiotemporal data fusion method, the Flexible Spatiotemporal DAta Fusion (FSDAF) method, to generate synthesized frequent high spatial resolution images through blending two types of data, i.e., frequent coarse spatial resolution data, such as that from MODIS, and less frequent high spatial resolution data such as that from Landsat. The proposed method is based on spectral unmixing analysis and a thin plate spline interpolator. Compared with existing spatiotemporal data fusion methods, it has the following strengths: (1) it needs minimum input data; (2) it is suitable for heterogeneous landscapes; and (3) it can predict both gradual change and land cover type change. Simulated data and real satellite images were used to test the performance of the proposed method. Its performance was compared with two very popular methods, the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and an unmixing-based data fusion (UBDF) method. Results show that the new method creates more accurate fused images and keeps more spatial detail than STARFM and UBDF. More importantly, it closely captures reflectance changes caused by land cover conversions, which is a big issue with current spatiotemporal data fusion methods. Because the proposed method uses simple principles and needs only one fine-resolution image as input, it has the potential to increase the availability of high-resolution time-series data that can support studies of rapid land surface dynamics.
    Full-text · Article · Jan 2016 · Remote Sensing of Environment
Show more