Skip Navigation

Journal of Plant Ecology 2008 1(2):137-141; doi:10.1093/jpe/rtn010
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chen, X.
Right arrow Articles by Zhang, X.-S.
Right arrow Search for Related Content
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2008. Published by Oxford University Press on behalf of the Institute of Botany, Chinese Academy of Sciences and the Botanical Society of China. All rights reserved. For permissions, please email: journals.permissions@oxfordjournals.org

Using spatial analysis to monitor tree diversity at a large scale: a case study in Northeast China Transect

Xiongwen Chen1,2,3,*, Bai-Lian Li2 and Xing-Shi Zhang3

1 Department of Natural Resources and Environmental Sciences, Center for Forestry, Ecology and Wildlife, Alabama A&M University, PO Box 1927, Normal, AL 35762, USA
2 Department of Botany and Plant Sciences, University of California, Riverside, CA 92521-0124, USA
3 Laboratory of Quantitative Vegetation Ecology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China

* Correspondence address. Center for Forestry, Ecology and Wildlife, Alabama A&M University, PO Box 1927, Normal, AL 35762, USA. Tel: +1-256-372-4231; Fax: +1-256-372-8404; E-mail: xiongwen.chen{at}aamu.edu


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Conclusions
 Supplementary Data
 References
 
Aims: Monitoring and assessing diversity change at a large scale is important for any meaningful biodiversity conservation and management. Spatial analysis techniques can provide information about different aspects of diversity distribution including change. We applied some common spatial analysis methods and additive partitioning of species diversity in the Northeast China Transect as a case study to show how to characterize the distribution and change of tree diversity in this area from different perspectives.

Methods: The field data were collected from the permanent plots conducted every 4 km. The additive partitioning of species diversity was used to characterize the diversity of tree species at different scales. Moran's I was used for identifying the spatial scale of autocorrelation, lacunarity was studied for diversity patch contagion and dispersion and spectral entropy was used for assessing the overall spatial distribution.

Important findings: Data collected from 1986 to 1994 indicate that the change of {alpha} diversity was not significant in the study area, but the change of β diversity was significant. The percentage of {alpha} diversity in total diversity ({gamma}) increased from 14.2 to 17.2%, and the percentage of β diversity decreased from 85.8 to 82.8%. For both {alpha} and β diversities, the scale of spatial autocorrelation decreased at the scale of 25–40 km and increased around 15–20 and 200 km. The lacunarity of {alpha} diversity decreased significantly and there was a sudden change at the scale of 56–68 km, but the lacunarity of β diversity increased across scales. The spectral entropy decreased slightly in {alpha} diversity and remained similar for β diversity. By using spatial analysis, we can monitor the diversity change over a large area and also assess the effectiveness of the current conservation strategies.

Keywords: monitoring • Northeast China Transect (NECT) • spatial analysis • tree diversity

Received: 27 August 2007 Revised: 6 March 2008 Accepted: 23 March 2008


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Conclusions
 Supplementary Data
 References
 
Monitoring biodiversity and characterizing biodiversity change at large scales are a fundamental step for understanding the underlying processes and preserving biodiversity at a regional level. Although humans can detect and respond best to local biodiversity change, persistent large scales forces are not so obvious. It is necessary to monitor and understand the spatial dimension of biodiversity distribution and possible underlying processes with respect to environmental change and human land use change. In addition, the spatial and temporal dimensions of biodiversity must be considered in experimental design and theory development. Diamond (1998) argued convincingly that biodiversity differs on a large scale due to climate, geography and biogeography at the various continents. The number and variety of statistical methods for spatial analysis of ecological information are developing. Simple methods of spatial analysis, such as spatial distribution or spatial autocorrelation, can detect change between different time periods and predict the possible effects of a specific process over a large area (Cliff and Ord 1981). For example, Cole and Syms (1999) applied spatial pattern analysis to distinguish causes of mortality of kelp in New Zealand. When using spatial analysis techniques, it is possible to obtain the general characteristics of biodiversity and monitor biodiversity change at a large scale. There are many methods of spatial analysis (e.g. Dale 1999; Fortin and Dale 2005; Haining 1990), but here as a case study we use some common techniques of spatial analysis with basic statistics to monitor tree species diversity in the forest area at Northeast China Transect (NECT).

NECT is identified as a middle-latitude transect for terrestrial ecosystem studies by the Global Change and Terrestrial Ecosystem, a core project of the International Geological and Biosphere Program (IGBP 1995) (see Fig. 1). Because this transect is parallel with latitude, the theoretical radiation is approximately uniform within the transect and its environmental gradient is mainly driven by moisture. In fact, the annual precipitation decreases from a high of 800 mm in the east to low of 100 mm in the west. Monitoring the change in the spatial pattern of tree diversity under the precipitation gradient will be helpful in the study of tree diversity dynamics in a large region and may lead to identifying the possible affecting processes. Some spatial characteristics and changes in tree diversity have been studied (Chen 2001; Chen et al. 2002), but using spatial analysis to assess the spatial change of tree diversity at a regional level is limited. It will be interesting to quantify the regional tree diversity change at NECT because the environmental gradient on NECT is sensitive to climate change and human disturbances. The aim of this study is to use the spatial change of tree diversity in the part of forest area of NECT from 1986 to 1994 as an example and apply some simple spatial analysis to detect the spatial change of tree diversity in a large area from different aspects of diversity pattern measurement.


Figure 1
View larger version (8K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1 The location of NECT in China. A colour version of this figure is available as supplementary data in the online version.

 

   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Conclusions
 Supplementary Data
 References
 
Study area
In this study, we chose the study area at the forest area of NECT although the whole NECT runs across forests, grasslands and deserts, the longitude from ~125°E to 131°E at a latitude of 43.55°N. The length of the study area is slightly >400 km. The data set was selected from the same permanent plot records conducted every 4 km in 1986 and 1994. The total plot number was 100. Each plot area was 30 x 30 m2 and all species were recorded from field surveys. We assumed that species composition in each plot could represent the species information of the corresponding area. The main tree species in this area are Acer mono, Betula platyphylla, Betula costata, Fraxinus mandshurica, Juglans mandshurica, Larix olgensis, Phellodendron amurense, Picea spp., Pinus koraiensis, Populus davidiana, Quercus mongolica, Tilia spp. and Ulmus pumila. Other information about this area can be found in Chen (2001) and Chen et al. (2002).

Methods
The additive partitioning of species diversity was used in this study. Lande's (1996) additive version treats {alpha} diversity as the average within-sample diversity, regardless of whether diversity is measured by species richness, the Shannon index or Simpson index; β diversity is the average amount of regional diversity not found in a single sample and {gamma} diversity is the total diversity at a study area. Generally, {gamma} diversity = average {alpha} diversity + β diversity. Because there is only one plot at each location in this study, here we used species richness at each location to represent {alpha} diversity and missing species number from regional species pool for β diversity. The additive partitioning of diversity has become more widely used because it allows for a direct comparison of {alpha} and β diversities at multiple spatial scales (here spatial scale means distance) and has potential applications in conservation (Veech et al. 2002).

The spatial scale of autocorrelation was computed using Moran's I with GS+TM5 (Gamma Design Software, Plainwell, MI, USA) for tree diversity in 1986 and 1994. The Moran's I statistic is a conventional measure of autocorrelation. With Moran's I, higher values indicate strong spatial correlation. In this study, the Moran's I is defined as the following (Robertson 2000):


Formula 1

(1)
where I is the measure of autocorrelation; n is the total number of samples; xi and xj are the observed values of diversity at site i and j, respectively; Formula is the average of x and S2 is variance. Variable wij is a symmetric weight matrix. In this study, wij is 1 if location j is within distance d from i or 0 otherwise. The maximum lag length is 200 km.

Lacunarity is an index that describes diversity patch contagion and dispersion at multiple scales (Plotnick et al. 1993). In this study, lacunarity is calculated as the variance-to-mean ratio ({Lambda} = variance/(mean)2 + 1) of the present tree species. Usually, they are presented in a double-log form: log({Lambda}(r)) as a function of log(r) and r is scale (Plotnick et al. 1996). Lower values indicate aggregation of habitats with higher richness of tree species ({alpha} diversity) at the given r scale. The gliding box procedure was repeated for each new box length (here it is equal to r) of 4, 8, 12, ... 100 km. The plot information of {alpha} diversity was used to represent corresponding {alpha} diversity at each box length.

The spectral entropy of diversity was estimated by the following (e.g. Crepeau and Isaacson, 1991; Li 2000):


Formula 2

(2)


Formula 3

(3)
where Ai is spectrum amplitude of the power spectral at frequency i by Fourier transformation. If the diversity is concentrated in a few frequencies, the spectral entropy will be relatively low. When the frequency of diversity is broadband, the spectral entropy becomes larger.

A paired t-test was used to test the difference in diversity between data collected in 1986 and 1994. The results are considered significant at P < 0.05.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Conclusions
 Supplementary Data
 References
 
Pattern of diversity distribution and percentage change of diversity
The {gamma} diversity was 18 in this study area. The general change in {alpha} diversity along the transect from 1986 to 1994 was not statistically significant (P > 0.05), but {alpha} diversity changed at many locations (Fig. 2a). The β diversity decreased significantly (P < 0.05) (Fig. 2b), and the change occurred mainly in the scale about 50–60 km.


Figure 2
View larger version (29K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 2 (a) {alpha} and (b) β tree diversities in the forest area on NECT in 1986 and 1994. A colour version of this figure is available as supplementary data in the online version.

 
For the entire study area, the percentage of {alpha} diversity in total diversity ({gamma}) increased from 14.2% in 1986 to 17.2% in 1994, and the percentage of β diversity decreased from 85.8 to 82.8%. For the whole study area, β diversity was more than {alpha} diversity. Using this method, we can obtain the percentage and its change of {alpha} and β diversities at any location in the study area to study their spatial variations. For example, the percentage of {alpha} diversity increased from 35.4 to 37.8% from 1986 to 1994 at the first 100 km in the study area.

Spatial autocorrelation
For both {alpha} and β diversities, the distance of the spatial autocorrelation increased slightly at the scale of 10–15 and 200 km and decreased around 20–40 km from 1986 to 1994 (Fig. 3). The change in the distance of spatial autocorrelation might indicate the increase or loss of tree species at this scale or some processes that can affect the tree species diversity operated at this scale.


Figure 3
View larger version (18K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 3 Spatial autocorrelation of {alpha} diversity in 1986 and 1994 (outside of the straight lines is 95% confidence area). A colour version of this figure is available as supplementary data in the online version.

 
Lacunarity
The lacunarity changes of {alpha} and β diversities from 1986 to 1994 were significant (P < 0.05) (Fig. 4). The lacunarity of {alpha} diversity increased but for β diversity decreased from 1986 to 1994 at all scales. With increased scales (length), {alpha} diversity increased along the transect from 1986 to 1994 but there were some ‘break points’ at log(r) 1.74–1.83 (the real scale length was approximately 56–68 km) where there were abrupt changes in {alpha} diversity. This kind of change occurred at the similar scales for β diversity.


Figure 4
View larger version (14K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 4 Lacunarity of {alpha} (a) and β (b) diversity in 1986 and 1994. A colour version of this figure is available as supplementary data in the online version.

 
Spectral entropy
The spectral entropy decreased slightly for {alpha} diversity (Fig. 5), and it remained similar for β diversity. This result indicated that the order or regularity of spatial distribution for {alpha} diversity increased.


Figure 5
View larger version (9K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 5 The spectral entropy of {alpha} and β diversities in 1986 and 1994. A colour version of this figure is available as supplementary data in the online version.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Conclusions
 Supplementary Data
 References
 
The β diversity was always linked to the notion of change along a gradient. In the context of the additive partitioning of species diversity, β diversity is used not only to measure the change along an environmental gradient but can also be used to quantify a change with a gradient. Veech et al. (2002) predicted a steady increase in the use of additive diversity partitioning, especially as a method for evaluating the statistical significance of diversity components. Diversity partitioning can improve biological surveys by identifying the primary sources of diversity in a region, e.g. see DeVries et al. (1997) and target conservation design (Fournier and Loreau 2001).

Although {alpha} diversity did not change significantly from 1986 to 1994, β diversity decreased significantly. The percentage of {alpha} diversity in total diversity ({gamma}) increased slightly, and the percentage of β diversity decreased slightly. There were spatial variations of {alpha} and β diversities in the study area, such as within the first 100 km in the study area where the percentage of {alpha} diversity increased. The change in diversity components could be due to ecological processes, such as intraspecific aggregation, habitat selection, limited dispersal capacity (Veech et al. 2002), precipitation fluctuation and local human disturbances (Chen 2001; Chen et al. 2002).

Spatial autocorrelation analysis was used to identify the spatial clustering (Ni et al. 2003) and operational units for conservation in continuous populations (Diniz-Filho and Telles 2002). The change of spatial autocorrelation reflects the spatial and temporal allocation of resources resulting from the interaction of species and their environment. In this study, for both {alpha} and β diversities, the distance of spatial autocorrelation increased slightly at 10–15 and 200 km and it decreased around 20–40 km. In such a way, spatial autocorrelation is useful to detect the change of spatial structure of diversity and find out the possible causes for such change.

By using lacunarity index, we can examine {alpha} and β diversity change across scales; it was evident that {alpha} diversity increased at different scales and there were some break points at the scale of 56–68 km (e.g. Fig. 4). Search for the possible causes of the break points may be useful to find the priority area for biodiversity conservation. The regularly spaced data of the same density produce curves that are initially straight and have lower lacunarity (Dale 2000). Lacunarity index was considered more versatile than indices that rely on a fixed spatial scale because the selection of scale could profoundly influence the results obtained and conclusions subsequently drawn (Riitter et al. 1996).

The overall spatial distribution was measured by the spectral entropy. The lower value means the higher order or regularity. In this study, it decreased slightly for {alpha} diversity, but it kept similarly for β diversity. This indicated that the order of spatial distribution for {alpha} diversity increased more than β diversity.


   Conclusions
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusions
Supplementary Data
 References
 
By using some commonly used spatial analysis techniques, tree diversity was found to change in the forest area on NECT from 1986 to 1994. The percentage of {alpha} diversity in total diversity ({gamma}) increased ~3%, and the percentage of β diversity decreased ~3%. For both {alpha} and β diversities, the distance of spatial positive autocorrelation increased at the scale of 10–15 and 200 km and decreased around 20–40 km from 1986 to 1994. The lacunarity of {alpha} diversity decreased significantly from 1986 to 1994 and there were sudden changes at the scale of 56–68 km and changes occurred for β diversity at the similar scale. The spectral entropy decreased slightly for {alpha} diversity. It is clear that spatial analysis can provide information of biodiversity in a large area from different aspects of measurement; it is useful for monitoring and managing diversity at large scale, such as finding the sensitive areas or priority areas and assessing the effectiveness of the current conservation strategies (e.g. Changbai Mountain Natural Reserve). Without doing spatial analysis, we may miss the big picture of diversity change and the possible underlying processes in a region. Without consideration of the spatial change of diversity, the conservation and management efforts at a large area will likely fail. Many methods of spatial analysis can be used to monitor diversity change at the regional level, especially with the help of high resolution remote sensing (e.g. Lidar products) to provide detailed information.


   Supplementary Data
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusions
 Supplementary Data
References
 
Supplementary data is available online at Journal of Plant Ecology.


   Acknowledgements
 
This work was partially supported by the University of California Agricultural Experimental Station and School of Agricultural and Environmental Sciences in Alabama A&M University. The authors thank Kathleen A. Roberts for editorial work.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusions
 Supplementary Data
 References
 

    Chen X. Change of tree diversity on Northeast China Transect (NECT). Biodivers Conserv (2001) 10:1087–96.[CrossRef]

    Chen X, Zhou G, Zhang X. Spatial characteristics and change for tree species along North East China Transect (NECT). Plant Ecol (2002) 164:65–74.

    Cliff AD, Ord JK. Spatial Processes: Models & Applications (1981) London: Pion Limited.

    Cole RG, Syms C. Using spatial pattern analysis to distinguish causes of mortality: an example from kelp in north-eastern New Zealand. J Ecol (1999) 87:963–72.[CrossRef]

    Crepeau JC, Isaacson LK. Spectral entropy as a measure of self-organization in transition flows. In: Self-Organization, Emerging Properties, and Learning—Babloyantz A, ed. (1991) New York: Plenum Press. 287–94.

    Dale MRT. Spatial Pattern Analysis in Plant Ecology (1999) Cambridge: Cambridge University Press.

    Dale MRT. Lacunarity analysis of spatial pattern: a comparison. Landsc Ecol (2000) 15:467–78.[CrossRef]

    DeVries PJ, Murray D, Lande R. Species diversity in vertical, horizontal, and temporal dimensions of a fruit-feeding butterfly community in an Ecuadorian rainforest. Biol J Linn Soc (1997) 62:343–64.[CrossRef][Web of Science]

    Diamond J. Guns, Germs and Steel. A Short History of Everybody for the Last 13000 Years (1998) London: Vintage.

    Diniz-Filho JA, Telles MPDC. Spatial autocorrelation analysis and the identification of operational units for conservation in continuous populations. Conserv Biol (2002) 16:924–35.[CrossRef][Web of Science]

    Fortin M-J, Dale MRT. Spatial Analysis: A Guide for Ecologist (2005) Cambridge: Cambridge University Press.

    Fournier E, Loreau M. Respective roles of recent hedges and forest patch remnants in the maintenance of ground-beetle (Coleoptera: Carabidae) diversity in an agricultural landscape. Landsc Ecol (2001) 16:17–32.[CrossRef]

    Haining R. Spatial Data Analysis in the Social and Environmental Sciences (1990) Cambridge: Cambridge University Press.

    IGBP. Spatial extrapolation and modeling on IGBP transects. Glob Change Rep (1995) 36:15–20.

    Lande R. Statistics and partitioning of species diversity, and similarity among multiple communities. Oikos (1996) 76:5–13.[CrossRef][Web of Science]

    Li BL. Why is the holistic approach becoming so important in landscape ecology? Landsc Urban Plan (2000) 50:27–41.[CrossRef]

    Ni S, Lockwood JA, Wei Y, et al. Spatial clustering of rangeland grasshoppers (Orthoptera: Acrididae) in the Qinghai Lake region of northwestern China. Agric Ecosyst Environ (2003) 95:61–68.[CrossRef]

    Plotnick RE, Gardner RH, Hargrove WW, et al. Lacunarity analysis: a general technique for the analysis of spatial patterns. Phys Rev E (1996) 53:5461–8.[CrossRef]

    Plotnick RE, Gardner RH, O'Neil RV. Lacunarity indices as measures of landscape texture. Landsc Ecol (1993) 8:201–11.[CrossRef]

    Riitter KH, O'Neil RV, Wickham JD, et al. A note on contagion indices for landscape analysis. Landsc Ecol (1996) 11:197–202.[CrossRef]

    Robertson GP. GS+: Geostatistics for the Environmental Sciences (2000) Plainwell, MI: Gamma Design Software.

    Veech JA, Summerville KS, Crist TO, et al. The additive partitioning of species diversity: recent revival of an old idea. Oikos (2002) 99:3–9.[CrossRef][Web of Science]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chen, X.
Right arrow Articles by Zhang, X.-S.
Right arrow Search for Related Content
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?