Summary Of Plant-Based Indicators Of Biodiversity And Soil Nutrient Availability
“Best bet” Land-use Systems
Thematic reports
Impact of different land uses on biodiversity
An Intensive Biodiversity Baseline Study in Jambi Province,Central Sumatra, Indonesia
Unique id: 11
Source file: D:\Projects\ASB\ASB Country and Thematic reports - xml\Above ground biodiversity assessmet WG\C-SEC-11.xml
Authors: A.N. Gillison
--------------------------------------------------------------------------
Introduction
At the time of writing newly acquired insect taxonomic and beetle trophic data have just come to hand. These will be added to the present data pool and the whole re-analysed to seek new and informative relationships between plant and animal taxa, functional types, and site physical variables. All the data sets thus far have been subjected to correlative analysis to identify those plant-based variables with highest predictive values along the land use intensity gradient. Because biodiversity is made up of interacting complexes of taxa, functional types, individuals and populations, correlation analysis is an appropriate first point of entry in seeking efficient predictors. High correlations do not necessarily indicate causal relationships although in some cases this can be reasonably inferred, especially where correlates bewteen different sets of variables vary consistently with soil nutrient availability.
Methods
Standard methods of regression (Pearson product-moment, using the MINITAB statistical package) were applied to help identify potentially useful plant-based surrogates of biodiversity. The sets of variables with highest linear correlates were also examined for non-linear pattern (second order polynomial regression as described in Part B, Annex II.) Because some of the highest correlates were associated with richness in plant species, modi and a ratio of species:modi, these variables were also examined for their potential value in predicting the occurrence of other taxa and functional groups as well as site physical variables. Table 11.1 below outlines the results.
Results
Plant species and PFTs or modi used alone and as a ratio tend to account for more variance in richness in animal taxa than animal taxa themselves. Overall there is a tendency for the species/modi ratio to greatly improve prediction for certain taxa and for above-ground carbon. This is evident in Table 11.1 a,b,c,d and is further illustrated in Annex II Figs 1a,b,c,d. for above-ground carbon, collembola, birds and termites. Table 11.2 lists correlations between plant-based variables and soil physico-chemical attributes.
11.4 Discussion
When the pattern of plant and animal taxonomic distribution along the LUTs is examined overall it is clear that the highest values tend to occur in certain pristine forest types and jungle rubber. This may be partly explained by the nature of the available niches in both intact forest and jungle rubber. Whereas the former has allowed the development of a series of cryptic terrestrial and arboreal habitats through a longer timespan, the relatively recent and more dynamic jungle rubber displays a much wider variety of niches due to the micro-fragmentary nature of the stand that is maintained by frequent disturbance by humans and animals. The cumulative graphs shown in Part B, Figure1.1 show considerable similarity between the area curves for BS05, the richest intact rain forest, and those for jungle rubber (BS10). The nature of the curves for the ratio values also relfects closely the general dynamic status of the LUT and whether it is degraded or not. The integration of the curves may provide a useful means for developing an index that reflects the overall indicator value of the species, modi and ratio combinations and this is the focus of a continuing study.
Of particular significance are the non-linear relationships of certain plant and vegetation structural variables with a sub-set of insect and bird taxa where the correlation is dramatically improved with ratios of species to PFT richness as distinct from correlations with either species or PFTs alone (Part B, Annex II, Fig. 1a,b,c,d) . There is no immediate explanation for this improvement although there is some suggestion of covariant patterns in soil nutrient availability. Table 11.2 outlines correlations between plant-based variables and a range of soil physico-chemical attributes. There are clear correlations between certain physical variables such as soil bulk density, soil pH, organic carbon and total N and Aluminium and species and PFT richness and increasing complexity in vegetation structure. The vegetation “V” index (ref: Part B) is also highly correlated with a range of soil variables. While diversity indices are rarely attributed much significance as biodiversity indicators, in the present study each of the Shannon-Wiener, Simpson’s and Fisher’s Alpha values are significantly correlated with a variety of key soil nutrients (Table 11.2).
11.5 Conclusions and recommendations
Before an appropriate synthesis can be made further investigation of more recently acquired data must be undertaken. However it is safe to say that the study so far clearly indicates that assessments of biodiversity should be designed to sample as much as possible of the spatial ranges of the taxa of concern to management. In general this translates to seeking out representative gradients of land use intensity and soil nutrient availability at landscape level and including climate at ecoregional level. Once these system boundaries have been located they offer a useful spatial and environmental context for identifying, calibrating and testing by spatial extrapolation, the best sets of biodiversity indicators. It is clear that there will be many situations where isolated samples taken for example from rain forest alone, are likely to give a misleading picture of regional patterns of biodiversity. The present study has revealed a new set of indicators that show promise for much wider application and testing. The study has shown that the plant-based attributes with taxonomic + functional complements possess potentially useful predictive value when coupled with certain taxa, soil nutrients and above-ground carbon. Taken together with readily measureable elements of vegetation structure such as mean canopy height and basal area, these offer an exciting prospect for examining the dynamics of biodiversity and associated land use at landscape scale.
Based on present findings, the message for managers of forested and agroforested lands is to maintain a mosaic of land cover types to maximise the availability of ecological niches. Not only is this likely to enhance biodiversity, recent experience suggests this may have a beneficial effect by facilitating biological pest management as well as providing increased flexibility for varying management options under conditions of environmental and socioeconomic change.
Table 11.1a Linear correlations
between beetle
trophic groups#, plant species and PFTs*
|
Troph.Grp |
PFT |
Species |
Spp/PFT |
|
Pchew1sp |
0.498 |
0.663 |
0.564 |
|
Pchew2sp |
0.581 |
0.585 |
0.369 |
|
Pchewspt |
0.539 |
0.650 |
0.508 |
|
Pchew1fm |
0.349 |
0.663 |
0.713 |
|
Pchew2fm |
0.764 |
0.590 |
0.190 |
|
Pchewfmt |
0.431 |
0.713 |
0.711 |
|
Pred1sp |
0.336 |
0.438 |
0.366 |
|
Pred2sp |
0.322 |
0.491 |
0.443 |
|
Predsptt |
0.354 |
0.467 |
0.394 |
|
Pred1fm |
0.608 |
0.449 |
0.115 |
|
Pred2fm |
0.044 |
0.324 |
0.448 |
|
Predfmtt |
0.542 |
0.521 |
0.286 |
|
Scav1sp |
0.411 |
0.404 |
0.245 |
|
Scav2sp |
0.407 |
0.755 |
0.781 |
|
Scavsptt |
0.437 |
0.503 |
0.372 |
|
Scav1fm |
0.165 |
0.362 |
0.395 |
|
Scav2fm |
0.338 |
0.794 |
0.928 |
|
Scavfmtt |
0.210 |
0.471 |
0.525 |
|
Tot1sp |
0.447 |
0.527 |
0.400 |
|
Tot2sp |
0.583 |
0.744 |
0.608 |
|
Utotalsp |
0.482 |
0.580 |
0.451 |
|
Totalsp |
0.401 |
0.337 |
0.156 |
|
Tot1fm |
0.335 |
0.577 |
0.572 |
|
Tot2fm |
0.450 |
0.854 |
0.899 |
|
Totfam |
0.339 |
0.573 |
0.562 |
# Pchew1sp = Phytophagous chewers – primary species;
Pchew2sp= secondary species;
Pchewt = total; -fm = family;
Pred = predators;
Scav = scavenger;
Utotalsp = total unique species.
*PFTs = Plant Functional Types;
Shaded areas with r = >0.500.
Bold type = high indicator value.
Table 11.1b Linear correlations
between beetle trophic groups
and vegetation structure*
|
Troph.Grp# |
|
Cr. Cov% |
|
Ba. Area |
|
Pchew1sp |
0.313 |
0.172 |
0.734 |
0.388 |
|
Pchew2sp |
0.248 |
0.232 |
0.529 |
0.329 |
|
Pchewspt |
0.298 |
0.197 |
0.679 |
0.376 |
|
Pchew1fm |
0.343 |
0.330 |
0.748 |
0.552 |
|
Pchew2fm |
0.600 |
0.572 |
0.222 |
0.530 |
|
Pchewfmt |
0.405 |
0.390 |
0.749 |
0.599 |
|
Pred1sp |
0.104 |
0.126 |
0.631 |
0.248 |
|
Pred2sp |
0.061 |
0.074 |
0.559 |
0.352 |
|
Predspt |
0.106 |
0.128 |
0.661 |
0.272 |
|
Pred1fm |
0.203 |
0.041 |
0.516 |
0.182 |
|
Pred2fm |
-0.090 |
-0.069 |
0.520 |
0.281 |
|
Predfmt |
0.138 |
0.007 |
0.660 |
0.273 |
|
Scav1sp |
-0.005 |
0.142 |
0.461 |
0.203 |
|
Scav2sp |
0.440 |
0.564 |
0.742 |
0.687 |
|
Scavspt |
0.088 |
0.239 |
0.550 |
0.317 |
|
Scav1fm |
-0.044 |
0.117 |
0.538 |
0.281 |
|
Scav2fm |
0.537 |
0.490 |
0.731 |
0.784 |
|
Scavfmt |
0.063 |
0.196 |
0.614 |
0.397 |
|
Tot1sp |
0.143 |
0.167 |
0.609 |
0.299 |
|
Tot2sp |
0.359 |
0.402 |
0.705 |
0.536 |
|
Utotalsp |
0.192 |
0.221 |
0.636 |
0.355 |
|
Totalsp |
-0.065 |
-0.008 |
0.478 |
0.103 |
|
Total1fm |
0.232 |
0.276 |
0.680 |
0.473 |
|
Total2fm |
0.572 |
0.549 |
0.712 |
0.846 |
|
Totalfm |
0.223 |
0.264 |
0.682 |
0.465 |
* Can.Ht = Mean canopy height (m)