Empirical Modelling of Vegetation Abundance from Airborne Hyperspectral Data for Upland Peatland Restoration Monitoring

Cole, Beth; Mcmorrow, Julia; Evans, Martin

doi:10.3390/rs6010716

Cited by 38 publications

(33 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Reflectance at and around the 970 nm absorption feature is related to vegetation moisture content and, in locations occupied by a high coverage of Sphagnum moss, can be an indicator of near-surface hydrology (Harris, Bryant, & Baird, 2005. A number of other studies have noted the importance of these spectral regions for identifying peatland PFTs or key PFT species (Bubier et al, 1997;Cole et al, 2014;Schaepman-Strub et al, 2009) and for mapping similar PFTs in tundra regions (Huemmrich et al, 2013).…”

Section: Discussionmentioning

confidence: 98%

“…There are very few studies that specifically attempt to directly map peatland PFTs from remote sensing data (Cole, McMorrow, & Evans, 2014;Schaepman-Strub et al, 2009;Schmidtlein, Feilhauer, & Bruelheide, 2012), primarily because of the difficulties involved in identifying characteristic spectral signatures using traditional spectralor pixel-based approaches. The main methods used are SMA (e.g., Schaepman-Strub et al, 2009), but difficulties occur in identifying representative and separable endmember spectra for each of the key PFTs; and regression modelling based on direct empirical relationships between PFT percentage coverage and reflectance, but such a method can prove to be problematic when mapping locations where the coverage of a PFT is low because of the presence of zeros in the regression models (Cole et al, 2014).…”

Section: Introductionmentioning

confidence: 99%

“…The main methods used are SMA (e.g., Schaepman-Strub et al, 2009), but difficulties occur in identifying representative and separable endmember spectra for each of the key PFTs; and regression modelling based on direct empirical relationships between PFT percentage coverage and reflectance, but such a method can prove to be problematic when mapping locations where the coverage of a PFT is low because of the presence of zeros in the regression models (Cole et al, 2014). Even though there are a limited number of PFTs commonly found within peatlands, we suggest that mapping PFT gradients may result in better models than those derived from single PFTs because gradient positions (i.e., where a plot lies along a floristic gradient) are inherently less variable than individual cover values (Schmidtlein & Sassin, 2004), especially in degraded peatlands or where restoration efforts are under way.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Hyperspectral remote sensing of peatland floristic gradients

Harris

Charnock

Lucas

2015

Remote Sensing of Environment

View full text Add to dashboard Cite

Previous studies have shown that the floristic composition of northern peatlands provides important information regarding ecosystem processes and their responses to environmental change. Remote sensing is the most expeditious method of obtaining floristic information at landscape and regional scales, but the spatial complexity of many northern peatlands and the spectral similarity of a number of peatland vegetation species is such that the success of traditional methods of vegetation classification is often limited. Here, we assessed whether ordination and regression analyses may be a useful alternative method for mapping peatland plant communities from remote sensing data. We used isometric feature mapping (Isomap) to describe the community structure of the peatland vegetation and related the identified continuous floristic gradients to hyperspectral imagery (AISA Eagle) using partial least squares regression (PLSR). We performed the same analysis at two hierarchical levels of species aggregation in order to map continuous gradients in the composition of both species and plant functional types (PFTs), the latter of which is the most widely used level of aggregation in northern ecosystems. Isomap was able to transfer 82% and more than 96% of the observed ground-based observations to the ordination space for plots characterised by species and PFT; respectively. The modelled floristic gradients showed good agreement with ground-based species and PFT observations although the strength of the agreement was proportional to the amount of floristic variation explained by each ordination axis (r2 val =0.74, 0.45 and 0.30 for the first three ordination axes and r2 val =0.68 and 0.66 for the first two ordination axes; for species and PFT floristic gradients respectively). We also found that how a PFT is defined has an important influence on the success with which it can be mapped. The resultant mapped floristic gradients enabled visualisation of homogeneous vegetation stands, heterogeneous mixtures of different key species and PFTs, and the presence of continuous and abrupt floristic transitions, without the need for unique spectral signatures or the collection of data characterising ancillary environmental variables.authorsversionPeer reviewe

show abstract

Section: Discussionmentioning

confidence: 98%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Hyperspectral remote sensing of peatland floristic gradients

Harris

Charnock

Lucas

2015

Remote Sensing of Environment

View full text Add to dashboard Cite

show abstract

“…This type of habitat mapping is based on remotely sensed data such as multispectral images [4], airborne hyperspectral imagery [5], light detection and ranging (LiDAR) [6,7], radar [8] and in some cases even a combination of these [9,10]. Nowadays, hyperspectral sensors have a high potential for monitoring of the environment [11][12][13][14].…”

Section: Introductionmentioning

confidence: 99%

Classification of Herbaceous Vegetation Using Airborne Hyperspectral Imagery

et al. 2015

View full text Add to dashboard Cite

Abstract:Alkali landscapes hold an extremely fine-scale mosaic of several vegetation types, thus it seems challenging to separate these classes by remote sensing. Our aim was to test the applicability of different image classification methods of hyperspectral data in this complex situation. To reach the highest classification accuracy, we tested traditional image classifiers (maximum likelihood classifier-MLC), machine learning algorithms (support vector machine-SVM, random forest-RF) and feature extraction (minimum noise fraction (MNF)-transformation) on training datasets of different sizes. Digital images were acquired from an AISA EAGLE II hyperspectral sensor of 128 contiguous bands (400-1000 nm), a spectral sampling of 5 nm bandwidth and a ground pixel size of 1 m. For the classification, we established twenty vegetation classes based on the dominant species, canopy height, and total vegetation cover. Image classification was applied to the original and MNF (minimum noise fraction) transformed dataset with various training sample sizes between 10 and 30 pixels. In order to select the optimal number of the transformed features, we applied SVM, RF and MLC classification to 2-15 MNF transformed bands. In the case of the original bands, SVM and RF classifiers provided high accuracy irrespective of the number of the training pixels. We found that SVM and RF produced the best accuracy when using the first nine MNF transformed bands; involving further features did not increase classification accuracy. SVM and RF provided high accuracies with the transformed bands, especially in the case of the OPEN ACCESS Remote Sens. 2015, 7 2047 aggregated groups. Even MLC provided high accuracy with 30 training pixels (80.78%), but the use of a smaller training dataset (10 training pixels) significantly reduced the accuracy of classification (52.56%). Our results suggest that in alkali landscapes, the application of SVM is a feasible solution, as it provided the highest accuracies compared to RF and MLC. SVM was not sensitive in the training sample size, which makes it an adequate tool when only a limited number of training pixels are available for some classes.

show abstract

“…Consequently, the field site has been the focus of recent research into carbon release, pollutant dynamics, and peatland restoration (e.g. Clay et al 2012;Dixon et al 2014;Cole et al 2014;Shuttleworth et al 2014Shuttleworth et al , 2015.…”

Section: Field Areamentioning

confidence: 99%

Contaminated sediment dynamics in peatland headwater catchments

et al. 2017

View full text Add to dashboard Cite

Abstract:Purpose: Despite substantial research into Pb storage in peatlands, formal description of the mechanisms of contaminated sediment mobilisation is limited. This study explores the controlling factors of contaminated sediment dynamics in an eroding peatland in the Peak District National Park, UK.Materials and methods: This study uses the Pb contamination stored near the peat's surface as a fingerprint to trace contaminated sediment dynamics in three severely degraded headwater catchments. A field portable XRF analyser was used across a range of catchment surfaces to examine patterns of contaminant storage and release.Results and discussion: Lead concentrations varied greatly over a small spatial scale. Erosion is exposing high concentrations of Pb on interfluve surfaces (up to 1660 µg g-1), and substantial amounts of reworked contaminated material (up to 1010 µg g-1) are stored on other catchment surfaces (gully walls and floors). A variety of factors have been shown to significantly control Pb release and storage in this environment, including wind action, aspect, and gully depth. Vegetation also plays an important role in retaining sediment-bound heavy metals within contaminated peat catchments.Conclusions: This study provides the first comprehensive overview of the mechanisms controlling Pb release and storage in degraded peatlands. Previous assessments of Pb fluxes may have underestimated contaminant export from severely degraded systems. Wind has also been identified as an as yet unaccounted for vector for heavy metal transport in peatland environments. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

show abstract

Empirical Modelling of Vegetation Abundance from Airborne Hyperspectral Data for Upland Peatland Restoration Monitoring

Cited by 38 publications

References 46 publications

Hyperspectral remote sensing of peatland floristic gradients

Hyperspectral remote sensing of peatland floristic gradients

Classification of Herbaceous Vegetation Using Airborne Hyperspectral Imagery

Contaminated sediment dynamics in peatland headwater catchments

Contact Info

Product

Resources

About