Notes towards Biodiversity Chapter 4
Introductory/Title Slide (1)
Hello. This is Gwen Raitt. I will be presenting this chapter on global biodiversity and its decline.
The pictures show the planets against the sun and the earth from space.
Oh Dear!
No one knows how much biodiversity there is or how much will be lost. The multiple levels of biodiversity mean that no single measurement for biodiversity is possible (Wikipedia Contributors 2006).
This chapter briefly considers some aspects of measuring ecosystem and genetic diversity before concentrating on the species inventory and estimates of global species numbers and species extinction rates. The present species inventory contains the 1.4—1.8 million species already described but does not contain much information about most of these species (Dobson 1996, Lovejoy 1997, Stork 1997).
Biodiversity at the Ecosystem Level
Inventorying ecosystems is complicated by the fact that it is difficult to set boundaries for an ecosystem. There is no standard classification system for ecosystems (Hawksworth and Kalin-Arroyo 1995). Ecosystems are usually classified at two levels: globally (with climatic determinants) and locally/regionally (with vegetation and species diversity determinants) (Hawksworth and Kalin-Arroyo 1995). The existing global classifications are inadequate (Bisby 1995).
Global ecosystem classification is of little value at the scales that are important for conservation (Hawksworth and Kalin-Arroyo 1995).
Remote sensing provides ways to assess and monitor vegetation structure and phenology – aspects of ecosystems (Hawksworth and Kalin-Arroyo 1995). The picture shows the Normalised Difference Vegetation Index (NDVI) for Africa for the months (starting at the top left) January, April, July and October.
Additional Notes
Extract from Wikipedia (2006A) ~ http://en.wikipedia.org/wiki/. “The Normalised Difference Vegetation Index (NDVI) is determined by measuring the ratio between visible and near infrared light reflected by the earth surface from satellites. One such satellite is employed by NOAA. It has a sensor called Advanced Very High Resolution Radiometer (AHVRR) which measures wavelengths of light ranging from 0.55–0.70 and 0.73–1.0 micrometers. If the reflecting surface is covered by green dense vegetation most of the visible light is absorbed and nearly all infrared light is reflected resulting in a high NDVI. The opposite happens if the vegetation is not green or the surface is only sparsely covered.
In general the equation for NDVI is: NDVI = (NIR – R)/(NIR + R)
Where R is the surface reflectance value for the red band and NIR is the surface reflectance value for the near infrared band.”
The Remote Sensing Course will provide more information.
Biodiversity at the Genetic Level
Genetic diversity may be considered/compared at three levels: the variability between individuals within a population, the variability between populations within a species and the diversity between species (Hawksworth and Kalin-Arroyo 1995).
Heterozygosity (the proportion of loci that carry two or more alleles) is used to quantify the variability between individuals within a population and the variability between populations within a species (Hawksworth and Kalin-Arroyo 1995). Comparisons of heterozygosity between species do not quantify how different the species are, merely how different their internal variability is. The picture shows heterozygous parents (both colours) and their homozygous (1 colour) and heterozygous offspring. The picture is simplified to show only 1 gene expressing a single trait.
The proportion of heterozygosity depends on: the evolutionary rates of the proteins or DNA used to measure the variability; the breeding system of the organism and the degree of connectivity between the populations (Hawksworth and Kalin-Arroyo 1995).
Problems with the Existing Species Inventory (1)
The precise number of recognised species is not known (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995, Stork 1997). There are several reasons for this.
One reason is the lack of a single definition for a species (Groombridge 1992, Bisby 1995, Hawksworth and Kalin-Arroyo 1995, Gaston 1996). Not all species definitions are comparable (Groombridge 1992, Gaston 1996). Using different species definitions to determine the number of species within a given taxon results in different total numbers of species for that taxon (Gaston 1996, Williamson 1997), e.g. using a species definition similar to that used by botanists on birds more than doubles the number of species identified (Williamson 1997).
Another reason is that the quality of the taxonomy is not consistent. The inventory contains poor taxonomy as well as sound taxonomy (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995).
The present inventory incorporates unknown amounts of synonymy. Synonymy refers to a single species being described and named more than once (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995, Stork 1997). Several factors contribute to the occurrence of synonymy. Firstly, there is no recognised central register of species though there are registers for some groups (Stork 1997). Secondly, holotype (‘type’) specimens may be difficult to access as they may occur in collections geographically far from the site at which they were discovered and traveling to view relevant collections is expensive – maybe prohibitively so. Finally, the natural variation of a new species is unknown – different forms may be given different names (Stork 1997).
Determining the number of recognised species at any time is not given priority, possibly because there is not much biological significance in the data (Groombridge 1992).
Additional Notes
Extract from Wikipedia (2006B) ~ http://en.wikipedia.org/wiki/. “A holotype is a single physical example (or illustration) of an organism” that is known to have been “used when the species (or lower-ranked taxon) was formally described. It is either the single such physical example (or illustration) or one of several such, but explicitly designated as the holotype.”
Problems with the Existing Species Inventory (2)
The existing partial inventory is biased towards: species that appeal to humans such as the giant panda (Ailuropoda melanoleuca – top picture); pests such as the cat flea (Ctenocephalides felis – bottom picture); organisms that do not require complex procedures or expensive equipment to study; larger size (the smaller the organism, the less likely it is to be studied); easily distinguishable species that are readily sorted and species that are easily accessed (Groombridge 1992).
Gaps in the Species Inventory
Gaps in the inventory may be considered in terms of the physical location and in terms of the categories of organisms. Usually the physical localities are expected to yield diversity in the categories of organisms that are not well inventoried.
Fitter et al. (2005) identified about 500 species of soil organisms (bacteria, protozoa and nematodes) at a single site in Scotland. This high alpha diversity will only lead to a large contribution to the global total if the species turnover (beta and gamma diversity) is also high.
The ‘hype’ about tropical forest canopies is not backed by clear evidence (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995). There is a possibility that the tropical forest floor has greater diversity than the canopy (Groombridge 1992, Stork 1997).
Marine benthic organisms show high alpha diversity but this will not lead to high global species totals unless the turnover of species (beta and gamma diversity) is also high. Many range sizes that are known are larger than those found on land and do not appear to support the idea of a high species turnover (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995). However, the macroscopic coral reef organisms of about 5—50 mm in length have a short dispersal distance (because their few larvae are relatively large and/or have shortened development periods) so their geographic ranges are restricted (Reaka-Kudla 1997) making it possible that the species turnover could be high.
The total richness of parasites and other symbionts will depend on the levels of host specificity (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995). Levels of parasitism depend on host size, host defenses and the population structure of the host. If the potential host is difficult to find, it will not have many obligate parasites (Groombridge 1992)
Nematodes may potentially be found as parasites, in marine sediments and in soils but little is known about species turnover (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995).
Fungi have high alpha diversity but there is insufficient information on latitudinal gradients and factors affecting turnover rates such as range size (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995). Species concepts are difficult to apply to microorganisms and there is little information on range sizes (Groombridge 1992). Perceptions of potential bacterial diversity have grown from studies suggesting that the numbers of unculturable bacteria are much greater than those of culturable bacteria (Hawksworth and Kalin-Arroyo 1995).
Plenty of evidence exists of high terrestrial arthropod diversity at most scales (the smallest scale is excluded). Insects form the largest portion of all described species (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995, Erwin 1997).
Clockwise from the top left, the pictures show a soil profile, the canopy of a rainforest, marine sediments, an insect larva covered in parasitic wasp pupae and a nematode worm.
Numbers (in Thousands) of Described Species
The numbers quoted here are from Groombridge (1992) and Hawksworth and Kalin-Arroyo (1995). Hawksworth and Kalin-Arroyo (1995) base their figures on Groombridge (1992) with updates. Please note that Groombridge (1992) lists vertebrates and Hawksworth and Kalin-Arroyo (1995) list chordates. Since vertebrates are a subset of chordates, the latter term was preferred. Differences in the numbers may reflect a combination of new species, discovered synonymies and the year to which the authors counted (Groombridge 1992, Hawksworth and Kalin-Arroyo 1995). All the numbers are in thousands of species (i.e. 4 = 4000 species).
Table 4.1: the numbers in thousands of described species from Groombridge (1992) and Hawksworth and Kalin-Arroyo (1995)
1992 / 1995
Viruses / 5 / 4
Bacteria / 4 / 4
Fungi / 70 / 72
Protozoans / 40 / 40
Algae / 40 / 40
Plants / 250 / 270
Chordates* / 45 / 45
Nematodes / 15 / 25
Molluscs / 70 / 70
Crustaceans / 40 / 40
Arachnids / 75 / 75
Insects / 950 / 950
Other / 96 + / 115
Total / 1700 + / 1750
Erwin’s Estimate of 30 million Arthropod Species
Erwin (1982) estimated that the global total of arthropod species was as large as 30 million. He based his estimate on a single study of the beetles collected by insecticide fogging the canopy of Luehea seemannii (a tropical evergreen tree) in Panama with some information from a study in Brazil (weevil numbers) (Erwin 1982, Groombridge 1992, Stork 1997). This estimate is used to demonstrate some of the problems with estimates based on single datasets.
The following ‘data’ were used to make the estimate. Three seasons of sampling nineteen Luehea seemannii trees yielded 955+ species of beetles excluding weevils (Stork 1997). From Brazil, he obtained the information that weevil numbers approximate leaf-beetle numbers so he suggested that there were about 206 weevils per tree species. Enquiry supplied the estimate of about 50 000 tropical tree species (Stork 1997).
Erwin recorded three assumptions. The first was that about 13.5% of the total number of beetle species per tree canopy were host specific. The second was that beetles make up 40% of the tropical canopy arthropods and the third was that the forest canopy to forest floor ratio is at least 2:1. Erwin actually added 1/3 of the number of canopy species to the total number of canopy species (Stork 1997). Erwin’s actual calculation thus assumed a tropical canopy to tropical floor arthropod species ratio of 3:1, though his stated assumption was at least 2:1.
Calculating Erwin’s Estimate
Add the 955 (beetle species minus weevil species) to the 206 (weevil species). This gives 1161 beetle species per tree species canopy. Round up to give 1200 beetle species per tree species canopy. If 13.5% of the beetle species per tree species canopy are host specific then there are 162 host specific beetle species per tree species canopy. The remaining 1038 beetle species are transient (Stork 1997). Multiplying the host specific beetle species per tree species canopy by the estimated number of tropical tree species yields 8.1 million host specific beetle species in the tropical forest canopy. Add the 1038 transient species and there are about 8 101 038 beetle species in the tropical forest canopy. If beetles make up 40% of all arthropods then there are about 20 252 595 arthropod species in the tropical forest canopy. Adding 1/3 of the total tropical forest canopy species to account for the tropical forest floor arthropods (Stork 1997) (note that this actually results in a 3:1 canopy to floor ratio) gives 26935951 tropical forest arthropod species. If the number of non-tropical arthropod species was estimated at about 3 million then the global total is approximately 30 million arthropod species.
More Recent Data and the Impacts on Erwin’s Estimate
Temperate and provisional tropical findings suggest that host specificity is less than 5%, thus invalidating Erwin’s assumption of 13.5% host specificity (Stork 1997). Though latitudinal variation in the proportions of species from different guilds of insects is probable, the figures from widely spread studies suggest that beetles make up 20—25% of the total number of arthropods, not the 40% assumed by Erwin (Stork 1997). Raw data from two studies suggest that the canopy to forest floor arthropod species ratio should at least be reversed (1:2 not 2:1 as Erwin suggested). There is also evidence to suggest that a large portion of the fauna will be found in both ecotones (Groombridge 1992, Stork 1997).
Using these figures and following the calculations shown before, the estimate of the global total for arthropod species becomes 39 013 680 arthropod species in the world. Changing the ratios used by Erwin makes a big impact on the total estimate (~39 million vs. ~30 million).
Problems with Single Sample Extrapolations
All estimates are affected by the accuracy of the figures used. The accuracy/completeness of counts from a single sample is open to question (Stork 1997). The calibration of ratios for single sample extrapolations is generally poor or non-existent. Erwin’s estimate is a case in point (Groombridge 1992).
There is an often unstated underlying assumption that the relationships used in the extrapolation scale evenly (Groombridge 1992) which is not necessarily true.
Useful Ratios for Estimation
All extrapolations from existing data involve one or more assumptions. The main one being that a ratio occurring in a known situation is also true in an unknown situation (Groombridge 1992).