Abstract
Remote sensing dynamic monitoring methods often benefit from a dense time series of observations. To enhance these time series, it is sometimes necessary to integrate data from multiple satellite systems. For more than 40 years, Landsat has provided the longest time record of space-based land surface observations, and the successful launch of the Landsat-8 Operational Land Imager (OLI) sensor in 2013 continues this tradition. However, the 16-day observation period of Landsat images has challenged the ability to measure subtle and transient changes like never before. The European Space Agency (ESA) launched the Sentinel-2A satellite in 2015. The satellite carries a Multispectral Instrument (MSI) sensor that provides a 10-20m spatial resolution data source providing an opportunity to complement the Landsat data record. The collection of Sentinel-2A MSI, Landsat-7 ETM+, and Landsat-8 OLI data provide multispectral global coverage from 10m to 30m with further reduced data revisit intervals. There are many differences between sensor data that need to be taken into account to use these data together reliably. The purpose of this study is to evaluate the potential of integrating surface reflectance data from Landsat-7, Landsat-8 and Sentinel-2 archived in the Google Earth Engine (GEE) cloud platform. To test and quantify the differences between these sensors, hundreds of thousands of surface reflectance data from sensor pairs were collected over China. In this study, some differences in the surface reflectance of the sensor pairs were identified, based upon which a cross-sensor conversion model was proposed, i.e., a suitable adjustment equation was fitted using an ordinary least squares (OLS) linear regression method to convert the Sentinel-2 reflectance values closer to the Landsat-7 or Landsat-8 values. The regression results show that the Sentinel MSI data are spectrally comparable to both types of Landsat image data, just as the Landsat sensors are comparable to each other. The root mean square error (RMSE) values between MSI and Landsat spectral values before coordinating the sensors ranged from 0.014 to 0.037, and the RMSE values between OLI and ETM + ranged from 0.019 to 0.039. After coordination, RMSE values between MSI and Landsat spectral values ranged from 0.011 to 0.026, and RMSD values between OLI and ETM + ranged from 0.013 to 0.034. The fitted adjustment equations were also compared to the HLS (Harmonized Landsat-8 Sentinel-2) global fitted equations (Sentinel-2 to Landsat-8) published by the National Aeronautics and Space Administration (NASA) and were found to be significantly different, increasing the likelihood that such adjustments would need to be fitted on a regional basis. This study believes that despite the differences in these datasets, it appears feasible to integrate these datasets by applying a linear regression correction between the bands.
Introduction
In Earth observation, "revisit time" is a critical factor for monitoring dynamic environmental changes. Tracking phenomena such as forest fires, crop phenology, or flood events requires a dense time series of observations. While the Landsat program has provided over four decades of data, its 16-day revisit cycle is often insufficient for capturing transient changes. The introduction of the European Space Agency’s Sentinel-2 Multispectral Instrument (MSI) offers a complementary data source with higher spatial resolution (10–20m) and more frequent revisits.
However, integrating Landsat-7 (ETM+), Landsat-8 (OLI), and Sentinel-2 (MSI) presents significant challenges. Each sensor possesses unique spectral response functions, spatial resolutions, and orbital characteristics. These hardware differences introduce systematic biases, making raw data stacks appear "noisy." Harmonization—the mathematical adjustment of spectral values to a common standard—is essential for reliable multi-sensor time-series analysis.
Methods and Materials
This study utilized the Google Earth Engine (GEE) cloud platform to process extensive datasets across China. The methodology involved five primary stages:
- Sampling: 5,000 random points were generated across China, representing diverse biomes including forests and croplands.
- Pairing: The researchers identified "near-simultaneous" observations where different sensors imaged the same location within a minimal time window to ensure ground conditions remained constant.
- Preprocessing: Sentinel-2 data were resampled from their native 10m/20m resolution to 30m to align with Landsat. Cloud and shadow masking were performed using the
s2cloudlessalgorithm. - Regression Modeling: Ordinary Least Squares (OLS) and Orthogonal Distance Regression (ODR) were employed to derive linear conversion coefficients for the Blue, Green, Red, NIR, SWIR1, and SWIR2 bands.
- Validation: The data were partitioned into training and validation sets to evaluate the robustness of the conversion models on independent observations.
Data Specifications
The study utilized "Tier 1" Surface Reflectance (SR) products. The spectral characteristics of the three sensors are compared in the table below, highlighting the variations in band centers (BC) and bandwidths (BW) that necessitate harmonization.
The resulting dataset included approximately 103,000 pairs for Landsat-7 vs. Sentinel-2, and 104,000 pairs each for Landsat-8 vs. Sentinel-2 and Landsat-7 vs. Landsat-8.
Results and Discussion
The harmonization process significantly improved spectral consistency across the sensors. Before adjustment, the Root Mean Square Error (RMSE) between Sentinel-2 and Landsat spectral values ranged from 0.014 to 0.037. Post-harmonization, these values were reduced to a range of 0.011 to 0.026. Similarly, the $R^2$ for the Normalized Difference Vegetation Index (NDVI) in the Sentinel-2 to Landsat-8 comparison improved from 0.642 to 0.751.
A key finding was the discrepancy between the regional coefficients derived for China and NASA’s global Harmonized Landsat Sentinel-2 (HLS) product. The significant differences observed suggest that localized land cover and atmospheric conditions can bias global models, underscoring the importance of regional harmonization for high-precision monitoring. Furthermore, the study found that Sentinel-2 MSI is spectrally more compatible with Landsat-8 OLI than with the older Landsat-7 ETM+ sensor.
Conclusion
While integrating data from the "Big Three" satellite systems is feasible through linear regression, several challenges remain. Future work must address Bidirectional Reflectance Distribution Function (BRDF) effects, residual atmospheric noise, and geometric co-registration errors. Nevertheless, this study demonstrates that regional harmonization effectively reduces spectral discrepancies, providing a more reliable foundation for dense time-series analysis in Earth observation.
Summary updated