Estimating natural grassland biomass by vegetation indices using Sentinel 2 remote sensing data

Filho, Marildo Guerini; Kuplich, Tatiana Mora; Quadros, Fernando Luíz Ferreira de

doi:10.1080/01431161.2019.1697004

Cited by 91 publications

(57 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…They provide a freely downloadable global coverage of the Earth's land surface every 10 days with one satellite and 5 days with two satellites. Given its good combination of spatial resolution and temporal frequency, S2 imagery is considered as having a great potential to improve pasture biomass assessment and monitoring [13][14][15][16] and has become one of the most popular remotely sensed sources in this research field. The high spatio-temporal resolutions of the S2 images are an important asset when monitoring pasture biomass in agricultural regions that are characterised by many small fields (~1 ha).…”

Section: Introductionmentioning

confidence: 99%

“…For example, Sibanda et al [14] reported that S2 optimally estimated biomass better than Landsat 8 OLI and performed somewhat comparable to hyperspectral bands. Filho et al [15] demonstrated that the S2 Multispectral Instrument (MSI) sensor on board the Sentinel-2A and Sentinel-2B satellites [16] provide quantitative indicators of the biomass status in natural grasslands with relatively good accuracy. Therefore, S2 data are likely to meet the challenge of providing accurate, regular biomass estimates in terms of two aspects: first, at a spatial resolution that is adequate for capturing the variations between typical-sized dairy paddocks, which may be as small as one hectare, and second, at a temporal resolution that is sufficient to detect the continuously changing landscape due to dairy cow rotations, which can be as frequent as every five days over different paddocks, and as frequent as less than 20 days in Spring in Tasmania.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Estimating Pasture Biomass Using Sentinel-2 Imagery and Machine Learning

et al. 2021

View full text Add to dashboard Cite

Effective dairy farm management requires the regular estimation and prediction of pasture biomass. This study explored the suitability of high spatio-temporal resolution Sentinel-2 imagery and the applicability of advanced machine learning techniques for estimating aboveground biomass at the paddock level in five dairy farms across northern Tasmania, Australia. A sequential neural network model was developed by integrating Sentinel-2 time-series data, weekly field biomass observations and daily climate variables from 2017 to 2018. Linear least-squares regression was employed for evaluating the results for model calibration and validation. Optimal model performance was realised with an R2 of ≈0.6, a root-mean-square error (RMSE) of ≈356 kg dry matter (DM)/ha and a mean absolute error (MAE) of 262 kg DM/ha. These performance markers indicated the results were within the variability of the pasture biomass measured in the field, and therefore represent a relatively high prediction accuracy. Sensitivity analysis further revealed what impact each farm’s in situ measurement, pasture management and grazing practices have on the model’s predictions. The study demonstrated the potential benefits and feasibility of improving biomass estimation in a cheap and rapid manner over traditional field measurement and commonly used remote-sensing methods. The proposed approach will help farmers and policymakers to estimate the amount of pasture present for optimising grazing management and improving decision-making regarding dairy farming.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Estimating Pasture Biomass Using Sentinel-2 Imagery and Machine Learning

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Remotely sensed data from unmanned aircraft systems (UAS) and satellites are one option to overcome this gap in the mapping and assessment of vegetation traits on a larger spatial scale 13 . In the last years, several studies made use of the potential of remotely sensed data to map grassland vegetation traits like above-ground biomass [14][15][16][17][18][19][20][21] and quality parameters [22][23][24][25][26][27] . Geo-referenced in-situ data are of utmost importance for the development of remote sensing-based models for the estimation of grassland vegetation traits, for both calibration and validation purposes.…”

Section: Background and Summarymentioning

confidence: 99%

Vegetation traits of pre-Alpine grasslands in southern Germany

Schucknecht

Krämer²,

Asam

et al. 2020

Sci Data

View full text Add to dashboard Cite

The data set contains information on aboveground vegetation traits of > 100 georeferenced locations within ten temperate pre-Alpine grassland plots in southern Germany. The grasslands were sampled in April 2018 for the following traits: bulk canopy height; weight of fresh and dry biomass; dry weight percentage of the plant functional types (PFT) non-green vegetation, legumes, non-leguminous forbs, and graminoids; total green area index (GAI) and PFT-specific GAI; plant water content; plant carbon and nitrogen content (community values and PFT-specific values); as well as leaf mass per area (LMA) of PFT. In addition, a species specific inventory of the plots was conducted in June 2020 and provides plot-level information on grassland type and plant species composition. The data set was obtained within the framework of the SUSALPS project (“Sustainable use of alpine and pre-alpine grassland soils in a changing climate”; https://www.susalps.de/) to provide in-situ data for the calibration and validation of remote sensing based models to estimate grassland traits.

show abstract

“…We calculated a series of spectral indices highlighting greenness, moisture and soil properties in order to increase utility of the spectral information contained in the original image bands (Table 3). Greenness, moisture and soil indices are derived from arithmetic combination of spectral information recorded in visible and near-infrared image bands and exhibit high correlation with vegetation characteristics such as phenology [52][53][54], biomass [55][56][57] and moisture content [58,59]. To complement the spectral information, spatial heterogeneity measures were calculated as a selection of simple and advanced Haralick texture features based on Gray Level Co-occurrence Matrix (GLCM) [60].…”

Section: Preparation Of Image Featuresmentioning

confidence: 99%

“…The coastal blue wavelength is absorbed by chlorophyll in healthy plants while the yellow band detects dryness/"yellowness" of vegetation, both of which are instrumental in vegetative analysis. The high importance of the spectral indices could be explained by their strong correlation with vegetation biomass [56,57] and moisture content [58,59], which helps to capture the varying characteristics of heterogeneous savannah vegetation that would otherwise be attenuated when using the original image bands alone. Additionally, the high importance of the texture features highlighted the chromatic variations in dry season savannah vegetation components.…”

Section: Spatial Patterns In Grazing Lawn Covermentioning

confidence: 99%

Probabilistic Mapping and Spatial Pattern Analysis of Grazing Lawns in Southern African Savannahs Using WorldView-3 Imagery and Machine Learning Techniques

et al. 2020

View full text Add to dashboard Cite

Savannah grazing lawns are a key food resource for large herbivores such as blue wildebeest (Connochaetes taurinus), hippopotamus (Hippopotamus amphibius) and white rhino (Ceratotherium simum), and impact herbivore densities, movement and recruitment rates. They also exert a strong influence on fire behaviour including frequency, intensity and spread. Thus, variation in grazing lawn cover can have a profound impact on broader savannah ecosystem dynamics. However, knowledge of their present cover and distribution is limited. Importantly, we lack a robust, broad-scale approach for detecting and monitoring grazing lawns, which is critical to enhancing understanding of the ecology of these vital grassland systems. We selected two sites in the Lower Sabie and Satara regions of Kruger National Park, South Africa with mesic and semiarid conditions, respectively. Using spectral and texture features derived from WorldView-3 imagery, we (i) parameterised and assessed the quality of Random Forest (RF), Support Vector Machines (SVM), Classification and Regression Trees (CART) and Multilayer Perceptron (MLP) models for general discrimination of plant functional types (PFTs) within a sub-area of the Lower Sabie landscape, and (ii) compared model performance for probabilistic mapping of grazing lawns in the broader Lower Sabie and Satara landscapes. Further, we used spatial metrics to analyse spatial patterns in grazing lawn distribution in both landscapes along a gradient of distance from waterbodies. All machine learning models achieved high F-scores (F1) and overall accuracy (OA) scores in general savannah PFTs classification, with RF (F1 = 95.73±0.004%, OA = 94.16±0.004%), SVM (F1 = 95.64±0.002%, OA = 94.02±0.002%) and MLP (F1 = 95.71±0.003%, OA = 94.27±0.003%) forming a cluster of the better performing models and marginally outperforming CART (F1 = 92.74±0.006%, OA = 90.93±0.003%). Grazing lawn detection accuracy followed a similar trend within the Lower Sabie landscape, with RF, SVM, MLP and CART achieving F-scores of 0.89, 0.93, 0.94 and 0.81, respectively. Transferring models to the Satara landscape however resulted in relatively lower but high grazing lawn detection accuracies across models (RF = 0.87, SVM = 0.88, MLP = 0.85 and CART = 0.75). Results from spatial pattern analysis revealed a relatively higher proportion of grazing lawn cover under semiarid savannah conditions (Satara) compared to the mesic savannah landscape (Lower Sabie). Additionally, the results show strong negative correlation between grazing lawn spatial structure (fractional cover, patch size and connectivity) and distance from waterbodies, with larger and contiguous grazing lawn patches occurring in close proximity to waterbodies in both landscapes. The proposed machine learning approach provides a novel and robust workflow for accurate and consistent landscape-scale monitoring of grazing lawns, while our findings and research outputs provide timely information critical for understanding habitat heterogeneity in southern African savannahs.

show abstract

Estimating natural grassland biomass by vegetation indices using Sentinel 2 remote sensing data

Cited by 91 publications

References 41 publications

Estimating Pasture Biomass Using Sentinel-2 Imagery and Machine Learning

Estimating Pasture Biomass Using Sentinel-2 Imagery and Machine Learning

Vegetation traits of pre-Alpine grasslands in southern Germany

Probabilistic Mapping and Spatial Pattern Analysis of Grazing Lawns in Southern African Savannahs Using WorldView-3 Imagery and Machine Learning Techniques

Contact Info

Product

Resources

About