Spectral distances explain higher variation in plant beta - diversity than spatial autocorrelation : abstract + powerpoint

Research output: Contribution to conferenceAbstractOther research output

Abstract

Most ecosystems lack long-term extensive data that can support analysis of what determines their inherent biodiversity patterns. The spectral variation hypothesis (SVH) suggests that between-plot differences in remotely sensed signal (spectral variation) is a proxy for environmental heterogeneity and therefore can act as an indicator of plant beta-diversity. Spatial process also controls plant betadiversity patterns. It is unclear how much of variation in beta-diversity that is explained by spectral variation can also be explained by spatial distances and vice versa. We evaluate the potential of spectral variation of Landsat-TM data and spatial autocorrelation in predicting plant beta-diversity in a fragmented landscape. We use Mantel correlograms to investigate the correlation between spectral distances and beta-diversity and the presence of spatial dependency in spectral distances and betadiversity. We fit a partial least squares regression (PLSR) to predict the variance in the plant betadiversity explained by spectral and spatial distances and partition the variances due to pure spectral, spatial dependency in spectral distances and purely spatial autocorrelation. Mantel correlograms
revealed an exponentially decaying correlation between beta-diversity and spectral distances that
was significant between plots with short spectral distances and at the farthest lag distances. Therefore dissimilarity in beta-diversity was lower at shorter spectral distances but increased with
increasing spectral distances. We detected significant spatial autocorrelation in beta-diversity and spectral distances suggesting that the variation in beta-diversity was influenced by spatial structure directly and indirectly through spatially dependent environment (spectral distances). The PLSR model
explained 37% of total variance in beta-diversity. Both environmental and spatial processes have significant control on beta-diversity patterns in this landscape though the former explained cumulative higher variance (66%) than spatial autocorrelation (34%). This suggests that conservation initiatives should aim at enhancing habitat diversity
Original languageEnglish
Pagess1-s16
Publication statusPublished - 11 Feb 2014
EventNetherlands Annual Ecology Meeting 2014 - Congrescentrum De Werelt, Lunteren, Netherlands
Duration: 11 Feb 201412 Feb 2014

Conference

ConferenceNetherlands Annual Ecology Meeting 2014
Abbreviated titleNAEM 2014
CountryNetherlands
CityLunteren
Period11/02/1412/02/14

Fingerprint

autocorrelation
Landsat thematic mapper
biodiversity
ecosystem
habitat

Keywords

  • METIS-302754

Cite this

@conference{9b6667e4687f444998d4b1ee6ad46e2f,
title = "Spectral distances explain higher variation in plant beta - diversity than spatial autocorrelation : abstract + powerpoint",
abstract = "Most ecosystems lack long-term extensive data that can support analysis of what determines their inherent biodiversity patterns. The spectral variation hypothesis (SVH) suggests that between-plot differences in remotely sensed signal (spectral variation) is a proxy for environmental heterogeneity and therefore can act as an indicator of plant beta-diversity. Spatial process also controls plant betadiversity patterns. It is unclear how much of variation in beta-diversity that is explained by spectral variation can also be explained by spatial distances and vice versa. We evaluate the potential of spectral variation of Landsat-TM data and spatial autocorrelation in predicting plant beta-diversity in a fragmented landscape. We use Mantel correlograms to investigate the correlation between spectral distances and beta-diversity and the presence of spatial dependency in spectral distances and betadiversity. We fit a partial least squares regression (PLSR) to predict the variance in the plant betadiversity explained by spectral and spatial distances and partition the variances due to pure spectral, spatial dependency in spectral distances and purely spatial autocorrelation. Mantel correlogramsrevealed an exponentially decaying correlation between beta-diversity and spectral distances thatwas significant between plots with short spectral distances and at the farthest lag distances. Therefore dissimilarity in beta-diversity was lower at shorter spectral distances but increased withincreasing spectral distances. We detected significant spatial autocorrelation in beta-diversity and spectral distances suggesting that the variation in beta-diversity was influenced by spatial structure directly and indirectly through spatially dependent environment (spectral distances). The PLSR modelexplained 37{\%} of total variance in beta-diversity. Both environmental and spatial processes have significant control on beta-diversity patterns in this landscape though the former explained cumulative higher variance (66{\%}) than spatial autocorrelation (34{\%}). This suggests that conservation initiatives should aim at enhancing habitat diversity",
keywords = "METIS-302754",
author = "K. Muthoni and T.A. Groen and A.K. Skidmore and A.G. Toxopeus",
note = "16 slides ; Netherlands Annual Ecology Meeting 2014, NAEM 2014 ; Conference date: 11-02-2014 Through 12-02-2014",
year = "2014",
month = "2",
day = "11",
language = "English",
pages = "s1--s16",

}

Muthoni, K, Groen, TA, Skidmore, AK & Toxopeus, AG 2014, 'Spectral distances explain higher variation in plant beta - diversity than spatial autocorrelation : abstract + powerpoint' Netherlands Annual Ecology Meeting 2014, Lunteren, Netherlands, 11/02/14 - 12/02/14, pp. s1-s16.

Spectral distances explain higher variation in plant beta - diversity than spatial autocorrelation : abstract + powerpoint. / Muthoni, K.; Groen, T.A.; Skidmore, A.K.; Toxopeus, A.G.

2014. s1-s16 Abstract from Netherlands Annual Ecology Meeting 2014, Lunteren, Netherlands.

Research output: Contribution to conferenceAbstractOther research output

TY - CONF

T1 - Spectral distances explain higher variation in plant beta - diversity than spatial autocorrelation : abstract + powerpoint

AU - Muthoni, K.

AU - Groen, T.A.

AU - Skidmore, A.K.

AU - Toxopeus, A.G.

N1 - 16 slides

PY - 2014/2/11

Y1 - 2014/2/11

N2 - Most ecosystems lack long-term extensive data that can support analysis of what determines their inherent biodiversity patterns. The spectral variation hypothesis (SVH) suggests that between-plot differences in remotely sensed signal (spectral variation) is a proxy for environmental heterogeneity and therefore can act as an indicator of plant beta-diversity. Spatial process also controls plant betadiversity patterns. It is unclear how much of variation in beta-diversity that is explained by spectral variation can also be explained by spatial distances and vice versa. We evaluate the potential of spectral variation of Landsat-TM data and spatial autocorrelation in predicting plant beta-diversity in a fragmented landscape. We use Mantel correlograms to investigate the correlation between spectral distances and beta-diversity and the presence of spatial dependency in spectral distances and betadiversity. We fit a partial least squares regression (PLSR) to predict the variance in the plant betadiversity explained by spectral and spatial distances and partition the variances due to pure spectral, spatial dependency in spectral distances and purely spatial autocorrelation. Mantel correlogramsrevealed an exponentially decaying correlation between beta-diversity and spectral distances thatwas significant between plots with short spectral distances and at the farthest lag distances. Therefore dissimilarity in beta-diversity was lower at shorter spectral distances but increased withincreasing spectral distances. We detected significant spatial autocorrelation in beta-diversity and spectral distances suggesting that the variation in beta-diversity was influenced by spatial structure directly and indirectly through spatially dependent environment (spectral distances). The PLSR modelexplained 37% of total variance in beta-diversity. Both environmental and spatial processes have significant control on beta-diversity patterns in this landscape though the former explained cumulative higher variance (66%) than spatial autocorrelation (34%). This suggests that conservation initiatives should aim at enhancing habitat diversity

AB - Most ecosystems lack long-term extensive data that can support analysis of what determines their inherent biodiversity patterns. The spectral variation hypothesis (SVH) suggests that between-plot differences in remotely sensed signal (spectral variation) is a proxy for environmental heterogeneity and therefore can act as an indicator of plant beta-diversity. Spatial process also controls plant betadiversity patterns. It is unclear how much of variation in beta-diversity that is explained by spectral variation can also be explained by spatial distances and vice versa. We evaluate the potential of spectral variation of Landsat-TM data and spatial autocorrelation in predicting plant beta-diversity in a fragmented landscape. We use Mantel correlograms to investigate the correlation between spectral distances and beta-diversity and the presence of spatial dependency in spectral distances and betadiversity. We fit a partial least squares regression (PLSR) to predict the variance in the plant betadiversity explained by spectral and spatial distances and partition the variances due to pure spectral, spatial dependency in spectral distances and purely spatial autocorrelation. Mantel correlogramsrevealed an exponentially decaying correlation between beta-diversity and spectral distances thatwas significant between plots with short spectral distances and at the farthest lag distances. Therefore dissimilarity in beta-diversity was lower at shorter spectral distances but increased withincreasing spectral distances. We detected significant spatial autocorrelation in beta-diversity and spectral distances suggesting that the variation in beta-diversity was influenced by spatial structure directly and indirectly through spatially dependent environment (spectral distances). The PLSR modelexplained 37% of total variance in beta-diversity. Both environmental and spatial processes have significant control on beta-diversity patterns in this landscape though the former explained cumulative higher variance (66%) than spatial autocorrelation (34%). This suggests that conservation initiatives should aim at enhancing habitat diversity

KW - METIS-302754

M3 - Abstract

SP - s1-s16

ER -