Machine learning methods’ performance in radiative transfer model inversion to retrieve plant traits from Sentinel-2 data of a mixed mountain forest

The Sentinel satellite fleet of the Copernicus Programme offers new potential to map and monitor plant traits at fine spatial and temporal resolutions. Among these traits, leaf area index (LAI) is a crucial indicator of vegetation growth and an essential variable in biodiversity studies. Numerous studies have shown that the radiative transfer approach has been a successful method to retrieve LAI from remote-sensing data. However, the suitability and adaptability of this approach largely depend on the type of remote-sensing data, vegetation cover and the ecosystem studied. Saltmarshes are important wetland ecosystems threatened by sea level rise among other human- and animal-induced changes. Therefore, monitoring their vegetation status is crucial for their conservation, yet few LAI assessments exist for these ecosystems. In this study, the retrieval of LAI in a saltmarsh ecosystem is examined using Sentinel-2 and RapidEye data through inversion of the PROSAIL radiative transfer model. Field measurements of LAI and some other plant traits were obtained during two succeeding field campaigns in July 2015 and 2016 on the saltmarsh of Schiermonnikoog, a barrier island of the Netherlands. RapidEye (2015) and Sentinel-2 (2016) data were acquired concurrent to the time of the field campaigns. The broadly employed PROSAIL model was inverted using two look-up tables (LUTs) generated in the spectral band’s settings of the two sensors and in which each contained 500,000 records. Different solutions from the LUTs, as well as, different Sentinel-2 spectral subsets were considered to examine the LAI retrieval. Our results showed that generally the LAI retrieved from Sentinel-2 had higher accuracy compared to RapidEye-retrieved LAI. Utilising the mean of the first 10 best solutions from the LUTs resulted in higher R2 (0.51 and 0.59) and lower normalised root means square error (NRMSE) (0.24 and 0.16) for both RapidEye and Sentinel-2 data respectively. Among different Sentinel-2 spectral subsets, the one comprised of the four near-infrared (NIR) and shortwave infrared (SWIR) spectral bands resulted in higher estimation accuracy (R2 = 0.44, NRMSE = 0.21) in comparison to using other studied spectral subsets. The results demonstrated the feasibility of broadband multispectral sensors, particularly Sentinel-2 for retrieval of LAI in the saltmarsh ecosystem via inversion of PROSAIL. Our results highlight the importance of proper parameterisation of radiative transfer models and capacity of Sentinel-2 spectral range and resolution, with impending high-quality global observation aptitude, for retrieval of plant traits at a global scale.

Download Full-text

Landsat-8 and Sentinel-2 Canopy Water Content Estimation in Croplands through Radiative Transfer Model Inversion

Remote Sensing ◽

10.3390/rs12172803 ◽

2020 ◽

Vol 12 (17) ◽

pp. 2803

Author(s):

Erik J. Boren ◽

Luigi Boschetti

Keyword(s):

Radiative Transfer ◽

Water Content ◽

A Priori ◽

Radiative Transfer Model ◽

Transfer Model ◽

Landsat 8 ◽

Multiple Sources ◽

Model Inversion ◽

Canopy Water ◽

Sentinel 2

Despite the potential implications of a cropland canopy water content (CCWC) thematic product, no global remotely sensed CCWC product is currently generated. The successful launch of the Landsat-8 Operational Land Imager (OLI) in 2012, Sentinel-2A Multispectral Instrument (MSI) in 2015, followed by Sentinel-2B in 2017, make possible the opportunity for CCWC estimation at a spatial and temporal scale that can meet the demands of potential operational users. In this study, we designed and tested a novel radiative transfer model (RTM) inversion technique to combine multiple sources of a priori data in a look-up table (LUT) for inverting the NASA Harmonized Landsat Sentinel-2 (HLS) product for CCWC estimation. This study directly builds on previous research for testing the constraint of the leaf parameter (Ns) in PROSPECT, by applying those constraints in PRO4SAIL in an agricultural setting where the variability of canopy parameters are relatively minimal. In total, 225 independent leaf measurements were used to train the LUTs, and 102 field data points were collected over the 2015–2017 growing seasons for validating the inversions. The results confirm increasing a priori information and regularization yielded the best performance for CCWC estimation. Despite the relatively low variable canopy conditions, the inclusion of Ns constraints did not improve the LUT inversion. Finally, the inversion of Sentinel-2 data outperformed the inversion of Landsat-8 in the HLS product. The method demonstrated ability for HLS inversion for CCWC estimation, resulting in the first HLS-based CCWC product generated through RTM inversion.

Download Full-text

Estimating crop primary productivity with Sentinel-2 and Landsat 8 using machine learning methods trained with radiative transfer simulations

Remote Sensing of Environment ◽

10.1016/j.rse.2019.03.002 ◽

2019 ◽

Vol 225 ◽

pp. 441-457 ◽

Cited By ~ 17

Author(s):

Aleksandra Wolanin ◽

Gustau Camps-Valls ◽

Luis Gómez-Chova ◽

Gonzalo Mateo-García ◽

Christiaan van der Tol ◽

...

Keyword(s):

Machine Learning ◽

Radiative Transfer ◽

Primary Productivity ◽

Landsat 8 ◽

Learning Methods ◽

Machine Learning Methods ◽

Sentinel 2

Download Full-text

Application of Landsat ETM+ and OLI Data for Foliage Fuel Load Monitoring Using Radiative Transfer Model and Machine Learning Method

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing ◽

10.1109/jstars.2021.3062073 ◽

2021 ◽

pp. 1-1

Author(s):

Xingwen Quan ◽

Yanxi Li ◽

Binbin He ◽

Geoffrey Cary ◽

Gengke Lai

Keyword(s):

Machine Learning ◽

Radiative Transfer ◽

Fuel Load ◽

Radiative Transfer Model ◽

Transfer Model ◽

Machine Learning Method ◽

Learning Method ◽

Load Monitoring

Download Full-text

Exploring continuous time series of vegetation hyperspectral reflectance and solar-induced fluorescence through radiative transfer model inversion

10.5194/egusphere-egu2020-14904 ◽

2020 ◽

Author(s):

Marco Celesti ◽

Khelvi Biriukova ◽

Petya K. E. Campbell ◽

Ilaria Cesana ◽

Sergio Cogliati ◽

...

Keyword(s):

Time Series ◽

Radiative Transfer ◽

Near Infrared ◽

Plant Traits ◽

Radiative Transfer Model ◽

Transfer Model ◽

Optimization Approach ◽

Ancillary Data ◽

Full Spectrum ◽

Retrieval Scheme

Remote sensing of solar-induced chlorophyll fluorescence (SIF) is of growing interest for the scientific community due to the inherent link of SIF with vegetation photosynthetic activity. An increasing number of in situ and airborne fluorescence spectrometers has been deployed worldwide to advance the understanding and usage of SIF for ecosystem studies. Particularly, a number of sites has been instrumented with the FloX (J&B Hyperspectral Devices, Germany), an automated instrument that houses two high resolution spectrometers covering the visible and near infrared spectral regions, one specifically optimized for fluorescence retrieval, the other for plant trait estimation.In this contribution we explore the feasibility to consistently retrieve plant traits and SIF from canopy level FloX measurements through the numerical inversion of a light version of the SCOPE model. The optimization approach was specifically adapted to work with the high- frequency time series produced by the FloX. In this context, a strategy for optimal retrieval of plant traits at daily scale is discussed, together with the implementation of an emulator of the radiative transfer model in the retrieval scheme. The retrieval strategy was applied to site measurements across Europe and the US that span a variety of natural and agricultural ecosystems.The full spectrum of canopy SIF, the fluorescence quantum efficiency, and main plant traits controlling light absorption and reabsorption were retrieved concurrently and evaluated over the growing season in comparison with site-specific ancillary data. Improvements and challenges of this method compared to other retrievals are discussed, together with the potential of applying a similar retrieval scheme to airborne datasets acquired with e.g. the HyPlant sensor, or the reconfigured &#8220;FLEX mode&#8221; data acquired with the recently launched Sentinel-3B during its commissioning phase.

Download Full-text