Optimal Exploitation of the Sentinel-2 Spectral Capabilities for Crop Leaf Area Index Mapping

Abstract

Abstract:The continuously increasing demand of accurate quantitative high quality information on land surface properties will be faced by a new generation of environmental Earth observation (EO) missions. One current example, associated with a high potential to contribute to those demands, is the multi-spectral ESA Sentinel-2 (S2) system. The present study focuses on the evaluation of spectral information content needed for crop leaf area index (LAI) mapping in view of the future sensors. Data from a field campaign were used to determine the optimal spectral sampling from available S2 bands applying inversion of a radiative transfer model (PROSAIL) with look-up table (LUT) and artificial neural network (ANN) approaches. Overall LAI estimation performance of the proposed LUT approach (LUT N50 ) was comparable in terms of retrieval performances with a tested and approved ANN method. Employing seven-and eight-band combinations, the LUT N50 approach obtained LAI RMSE of 0.53 and normalized LAI RMSE of 0.12, which was comparable to the results of the ANN. However, the LUT N50 method showed a higher robustness and insensitivity to different band settings. Most frequently selected wavebands were located in near infrared and red edge spectral regions. In conclusion, our results emphasize the potential benefits of the Sentinel-2 mission for agricultural applications.

Toward Generic Models for Green LAI Estimation in Maize and Soybean: Satellite Observations

“…Given that these crops have different photosynthetic pathways (C3 vs. C4), leaf structures (dicot vs. monocot) and canopy architectures (a heliotropic leaf angle distribution vs. a spherical leaf angle distribution), they cover a large spectrum of biophysical conditions. It would be of great interest for a further study to compare LAI estimation by the models presented here and by, e.g., a NN model based on data simulated by radiative transfer models (e.g., [12,19]). Such a study would give important insights into the realism of the pigmentation, leaf reflectance, and canopy structural properties that are assumed by such radiative transfer models.…”

“…It is also consistent with the results of Feilhauer et al [59] using a multi-method ensemble that provides a robust selection of spectral bands related to biochemical traits obtained for foliar chlorophyll content retrieval. Richter et al [12,20] showed, using simulated data taken at close range, that MSI NIR and red edge spectral bands provided the most relevant information for LAI estimates. Weiss et al [55], Bacour et al [19], and Baret and Buis [2] came to the same conclusion.…”

“…However, several studies have shown that increasing the number of bands above three or four did decrease the NRMSE of models trained and validated by simulated or aircraft data-seven spectral bands were needed to achieve the minimal error by NN using aircraft data [54], four to six (NN) and five to six (look up tables) band combinations using Thus, three or four bands were sufficient to minimize the NRMSE of LAI estimation using satellite data by both Sim and Sat models. However, several studies have shown that increasing the number of bands above three or four did decrease the NRMSE of models trained and validated by simulated or aircraft data-seven spectral bands were needed to achieve the minimal error by NN using aircraft data [54], four to six (NN) and five to six (look up tables) band combinations using simulated data and data taken at close range [12], and six bands were selected by NN from nine synthetic bands [55]. Richter et al [12] noted that while the results of selection analyses may depend on the algorithms employed, generally, the optimal number of bands for estimation of LAI is around six to eight.…”

“…However, several studies have shown that increasing the number of bands above three or four did decrease the NRMSE of models trained and validated by simulated or aircraft data-seven spectral bands were needed to achieve the minimal error by NN using aircraft data [54], four to six (NN) and five to six (look up tables) band combinations using simulated data and data taken at close range [12], and six bands were selected by NN from nine synthetic bands [55]. Richter et al [12] noted that while the results of selection analyses may depend on the algorithms employed, generally, the optimal number of bands for estimation of LAI is around six to eight. Moreover, they showed that the worst results were achieved through look up tables in combination with a two-band arrangement.…”

Sentinel-2 for Mapping Iron Absorption Feature Parameters

“…Modified after [2,3] and used with permission from [4]. The potential of Sentinel-2 has been explored for a variety of land surface parameter estimations: Leaf Area Index (LAI) [5][6][7]; chlorophyll and nitrogen [8]; biophysical parameters [9,10]; and red edge position [8,11]. A study on Normalized Difference Vegetation Index (NDVI) time series continuity was done for operational sensors, which included National Oceanic and Atmospheric [12].…”