Citation: Tokumine, S. 2002. Bioclimatic modelling of sub-Saharan African plants. Progress report to Conservation International, September 2002. Centre for Ecology Law and Policy, Environment Department, University of York, UK.
Global climate is changing, and is projected to continue changing (Hughes, 2000). At large scales of 10 km2 and above, macroclimate has been seen as a crucial element in the distribution patterns of many organisms (Huntley et al. 1995). The rate of genetic adaptation is unlikely to match the speed of climate change (Etterson & Shaw 2001). This change is already believed to have had an impact on many natural systems (IPCC 2001a), or is predicted to cause major changes to biodiversity for which new conservation paradigms must be established (Peterson et al. 2002; Hannah et al. 2002).
These new paradigms need predictions of potential future change on which to base current conservation strategy. Ideally, this information would be provided by long-term, controlled lab experiments monitoring the direct effects of changing climatic variables (e.g. Davis et al. 1998). Ideally, these lab experiments would then be combined with extensive field experiments in all the major vegetation types of the world, so nurturing a deep understanding of the processes underlying the community structure in all cases (Johnston & Schmitz, 1997). It’s unfortunate that this knowledge is likely to remain unavailable in the timeframe for dramatic climate change predicted by the IPCC (Root & Schneider, 1993; IPCC, 2001a).
One commonly used alternative approach to predict the effect on the distribution of biota under climate change is based on the assumption that species distributions are directly dependant on local climate (Walter, 1979). At its simplest it involves “linking” species’ current distributions with combinations of current climate data, then plotting spatial shifts of these “climate envelopes” using climate change scenario data. Global geological histories provide clues as to the possible effects of climate change on the biota of the earth (Graham & Grimm, 1990; Webb, 1987). Changes in the distribution patterns of these terrestrial organisms have been shown to be the primary response to paleological climate changes with macroevolutionary response to climatic fluctuations appearing limited in comparison (Huntley & Webb, 1989). Though we cannot be certain past changes are directly comparable to the changes we will experience due to the differing combination of causal factors (Graham & Grimm, 1990), we can expect the distributions of current plant species to exhibit a distributional response with future changes in global climate. The linking approach has come under criticism for not incorporating the possible impacts of changing inter and intra-species interactions under different climate change scenarios (Davis et al., 1998). Though acquisition of this knowledge is important, its absence is symptomatic of the lack of experimentation highlighted above. “Linking” techniques are therefore contemporarily important as ‘null models’ from which we can view the possible changes due to global warming.
Previous methods of implementing the linking approach are broadly based on two methods, generalised linear models and BIOCLIM (Nix, 1986) approaches. Generalised linear models have been mainly used in instances where there are largely complete datasets incorporating absence data (Yee & Mitchell 1991). An extension of this has seen the use of climate response surfaces (Bartlein et al. 1986; Hill, Thomas & Huntley, 1999). BIOCLIM-type approaches have been used on less complete datasets which focus on species presences (Busby, 1986; Nix, 1986, Eeley et al.1999). Combinations of the two approaches have also been proposed under the Genetic Algorithm for Rule Set Prediction (GARP) project, which uses a combination of BIOCLIM rules, logistic regression and machine learning methods. (Peterson et al. 2002).
BIOCLIM type approaches are based around evaluating a species’ “Climate Envelope”. These are areas defined by the overlay of a number of ranges of climate variables. These ranges describe the minimum and maximum values of a climate variable found at the location where a species occurrence is recorded. In this way, all areas exhibiting a combination of climatic conditions within the range of conditions dictated by a species’ distribution are found. This method therefore delineates climatically suitable areas for the species, in other words, their “climate envelope”. If evaluating a species using two climate variables, annual rainfall and temperature for example, we may find the range of rainfall and temperature experienced across all points in the species distribution to be between 150-250 mm and 10-15oC respectively. All areas within the study area that satisfy both these requirements would make up this species’ climate envelope. As more climate variables are added, the description of suitable climate becomes increasingly specific to the species distribution, resulting in a climate envelope more spatially representative of that species distribution. However, there are limits to the extent to which this process is practicable. Using typically available sets of relevant climate variables, the overlay technique of BIOCLIM can often result in over-predictions of suitable areas.
Removal of a point in a given species distribution can result in a more narrowly defined climate envelope, potentially leading to a reduction in these overpredictions. Minimising overpredictions while maximising the climate envelopes capture of actual species occurrences becomes a balancing act. A number of indices exist that quantify this trade-off. These indices are functions of the number of correctly and over or under predicted areas, addressing issues of reliability important in many areas of GIS and remote sensing (Story & Congalton, 1986). Though the Kappa statistic has been suggested as an index of fit which most reliably represents the fit of a climate envelope to a given species original distribution (Manel et al., 2001), the index requires absence data often lacking from library species distribution records. In order to assess the quality of a climate envelope in spatially delineating a given species’ distribution where only presence data is available, a normalising similarity indices such as Sørenson’s Index can be used (Clifford & Stephenson, 1975).
To find the climate envelope for a given set of climate variables with the best possible fit to the actual species distribution according to Sørenson’s Index , we could construct a climate envelope and calculate the Sørenson’s Index for all possible combinations of points of a species distribution. For small ranging species, all possible combinations can easily be computed. However, to assess the climate envelopes for all combinations of a species distribution comprising, for example, 50 points would result in a number of possible combinations so large the feasibility of this “brute force” approach is limited.
The problem of finding this optimal index score in a field of possible solutions is a classic example of a “NP – complete” Boolean Satisfiability Problem (SAT); a class of computational problems for which no efficient solution algorithm has been found (DeJong & Spears, 1989). When an NP – complete problem must be solved, one approach is to use computational heuristics to obtain an optimal or near-optimal solution. NP – complete SAT’s have been shown to be effectively solved by genetic algorithms (GAs) (DeJong & Spears, 1989). GAs are adaptive heuristic search algorithms based on the idea of natural selection. Initially developed by Holland (1975), they have been widely applied in many fields where there are NP-complete problems from fields as diverse as music and circuit board design. GAs have also been used in spatial analysis. Applications range from the definition of catchment areas for building society branches (Hobbs 1995) to optimal patch configuration (Brookes, 2001). One approach to defining climate envelopes using a GA to develop decision rules is also reported by Stockwell and Peters (1999).
In addition to providing a method to find the near optimal climate envelope, the heuristic optimisation approach of GAs has certain advantages over more traditional statistical approaches to creating a predictive model of species presence/absence. Notably, possible regression techniques such as logistic regression may be affected by over-dispersion caused by model miss-specification. This can result from the spatial autocorrelation found in the climatic independent variables and the dependent variable itself. It also may make intuitive ecological sense to consider ranges of climate variable values that may be suitable for the occurrence of a species, rather than using statistical approaches implying single, optimal, variable values associated with species area occurrence.
Once defined, by substituting the current, observed environmental variables used to define the climatic envelope with those derived from future atmospheric general circulation models such as HadCM3 (Gordon et al. 2000), potential spatial changes in a species distribution are elucidated. Standard GIS software is used for the manipulation of the present and future climate surfaces, along with the species spatial distributions considered. Once analysed by the GA, results from the consideration of these surfaces are returned to the GIS for mapping and the calculation of overlap between observed and predicted species distributions. The GIS also allows an assessment of the spatial concurrence of present observed species distributions and future predictions under climate change.
We describe a tested, GA based methodology for the modelling of species distributions in climate space for which only presence data is available. We also describe the development of Tanzanian Eastern Arc tree species datasets, Tanzanian DEM aggregation results, present and future climate and soil surface datasets at the continental scale and emergent issues from this. These components are discussed within the context of high resolution Tanzania scale modelling of 452 tree taxa and low resolution continental scale climate envelope modelling of over 3500 plant species.
1. African continent scale climate change surfaces for the present day, 2025, 2055, 2085.
2. High resolution Tanzanian Digital Elevation Models (DEM) and soil datasets.
3. High resolution Tanzanian Tree species dataset consisting of 452 tree taxa at 220 locations across the Eastern Arc Mountains.
4. Genetic algorithm software program development complete.
5. Linking and automation program between the GA and WORLDMAP plant mapping software development complete.
6. 3566 individual genetic algorithm optimised climate envelopes constructed for the entire sub-Saharan Africa continental WORLDMAP plant database (~10% of total African flora). Allowing accurate reproduction of species richness and endemism using only climate variables (Figure 1 & 2).
7. Projection of all 3566 climate envelopes to climate predicted for 2025, 2055 & 2085, together with per species information on extinction, colonisation, and species turnover (table 4).
8. Genetic Algorithm methodology submitted to the journal of Computers Environment and Urban Systems.
9. Genetic Algorithm methodology presented at the GIS researchers UK conference, the British Ecological Society macroecology conference, and at a workshop hosted by this project at the University of York.
The GA starts with a “population” of possible climate envelopes defined by a set of climate variable ranges. Each of these possible solutions is described as a chromosome. For example, one chromosome from the population might have a minimum value for mean maximum temperature of 25 deg. C, a maximum value for mean maximum temperature of 32 deg. C, a minimum value for mean annual precipitation of 600 mm, a maximum value for mean annual rainfall of 850 mm and so on, until ranges are defined for each of the climate variables included in the analysis. In this case, each minimum or maximum value represents one “gene” on a chromosome. The maximum and minimum values for the variable ranges, the genes, are chosen randomly from within the range of the variables found in the entire study area. In addition to the randomly created chromosomes, a chromosome formed by the climate ranges contained within the spatial range of the species in question is included along with a chromosome containing the climate ranges for the entire study area.
The areas accounted for by the set of variable ranges contained in the genes of a chromosome are then mapped and compared to the species distribution being modelled by using the following version of Sørenson’s Index:
S = 2a/2a + b +c
Where S is the Sørenson’s Index score, a is the number of correctly predicted occurrences, b is the number of falsely predicted occurrences and c is the number of unpredicted occurrences. This gives a normalised score, or fitness value, of how well the envelope given by the variable ranges of the chromosome accounts for the spatial distribution of the species.
On the basis of their fitness values, members of the population are allowed to reproduce via a crossover operation. That is, if a randomly chosen chromosome has a high fitness score it has a higher probability of reproducing with another randomly chosen chromosome from the population. In this way “good genes” are taken from successful individuals into the next generation in a search for the optimal solution.
In an attempt to avoid finding local maxima resulting from the limited nature of the genes of the initial population an operation known as mutation is added to the algorithm. The mutation operation simply allows the value of a randomly selected gene on a randomly selected chromosome to be increased or decreased by a random number drawn from a normal distribution, N(0,2). The algorithm advances through generations of solutions until no noticeable improvements are made to the fit of the climate envelope to the species distribution.
Testing the Genetic Algorithm
To compare the efficacy of the GA compared to the more common Logistic Regression method, the different approaches were applied to two test systems representing “hard” and “easy” searches on grids with known characteristics based on a DEM of Madagascar derived from the GTOPO30 DEM (USGS 1999). The easy test system consisted of 3 DEM derived variables: elevation, slope and hillshade. A “dummy species” distribution representing 20% of all cells was created using subsets of the variable ranges. The hard test system consisted of 5 DEM derived variables: elevation, slope, hillshade, profile and planform curvature. In this case the dummy species distribution represents 1% of all cells in the test system. The optimal solutions for each system were therefore known and attainable. For the logistic regression tests, a surface representing probability of species occurrence on a cell-by-cell basis was constructed. To derive the probability threshold of the logistic regression output which best represented the dummy species distribution, a method of iteratively evaluating the Sørenson’s Index in steps of 0.05 for all probabilities was used in order to find the threshold that yields the maximum Sørenson’s Index score (Huntley et al. 1995).
In the easy test system, logistic regression gives comparable envelope results to the GA methodology. However in the hard test system it is clear that the GA outperforms the logistic regression based approach (Table 1).
TABLE 1. Numbers of original and modelled occurrences with Sorensen’s index of similarity scores (1 = perfect fit)
+ “Easy” pseudo-species constructed from altitude, shade and profile derivatives
++ “Hard” pseudo-species constructed from altitude, shade, profile, profile curvature and planform curvature derivatives
* p >= 0.45
** p >= 0.1
In this instance, the logistic regression technique appeared to predict well, with an overall correct prediction of 96.9% of cells. However, though an overall high percentage was correctly predicted, the majority of cells correctly predicted are the more abundant cells considered in a logistic regression framework as being absence cells. The technique failed to predict 47% of the less abundant “species” presence values.
The hard test system was also used to investigate the effect of varying initial GA mutation and chromosome number parameters on final climate envelope fit and computational time. As GAs are heuristic, each parameter variant was tested for consistency via repetition.
In initial testing, as expected, lack of a mutation step was found to result in a low final fitness score, as the GA is unable to move away from local maxima. The addition of 25 or more mutation steps results in a near-optimal solution being found. Though the starting population can be made up of any number of individual chromosomes, 100 individuals produce consistent results. The effect on computational time of adding both chromosome and mutation steps to the algorithm appears to be linear (R2 = 0.9944, Time = 0.0416 x number of chromosomes + 9.6773; R2 = 0.993, Time = 0.0802 x number of mutations + 3.6667).
Dataset aggregation and development
This project illustrates the modelling of both high and low resolution plant species datasets. To accomplish this, digitised surfaces of both high and low resolution present and future climate and other environmental parameters are required for each modelling step.
Sub Saharan Africa low resolution plant species dataset
Low resolution plant taxa distributional data used in this analysis were taken from a WORLDMAP (Williams, 1992) based database of 3566 sub-Saharan plant species mapped at 1-degree resolution (Appendix A). This gives a cumulative total of 54469 individual points of species presence. The plant taxa are recorded on a 1-degree grid representing continental sub-Saharan Africa in 2075 cells extending across the latitudinal extent of sub-Saharan Africa from the Cape of Good Hope at 34.5 S to 19.5 N and are subject to post-entry data verification techniques (Lovett et al. 2000; La Ferla et al. 2002; Bürger, in press.). These have been collated at York as part of an ongoing Conservation International project established in 1996 to digitise African plant distributions (CELP, 2001). This data entry was also conducted in collaboration with the Botanical Museum, University of Copenhagen. The data in the database have been assimilated from numerous public domain sources including plant taxonomic revisions, monographs, and personal communication with collectors, and currentlyrepresents the single most extensive systematic mapping of continental sub-Saharan African higher plants. Conversion algorithms were constructed to extract species distributional data from the database in a format recognised by the GA. Work on improving the plant data is ongoing in collaboration with the Biomap Project at Bonn.
Tanzanian high resolution plant datasets
High resolution plant species distributional data allows us to better understand the processes operating on species distributions at the local and regional scale. Furthermore, it corroborates and verifies trends and patterns of species distributions derived from the lower resolution 1 degree African plant mapping, especially in the case of restricted range species. To develop an expert verified plant species GIS database for this purpose, the locations of 452 Tanzanian tree taxa were precisely geo-referenced at 220 locations and recorded in a point data format. This has resulted in a dataset containing a total of 4321 records & the data set is still being expanded. These data cover many areas of the Tanzanian Eastern Arc Mountains, and represent the digitisation of field work by J. Lovett.
High resolution soil dataset
The soils dataset available for this study comes from the Reynolds et al. (1999) 0.083 decimal degree soils dataset of estimated available water content from the FAO soil map of the world, global soil profile databases, and pedo-transfer functions. The dataset specifically contains: particle-size distribution; dominant soil texture; organic carbon content; coarse fragments; bulk density, porosity, available water content, Soil water-holding capacities for two-layers of depths (0-30 and 30-100 cm) and soil depth for two layers of maximum depths (100 cm and 250 cm). This dataset is well suited to analysis of this kind due to it’s continental coverage. Although the Soil and Terrain Database (SOTER) provides higher resolution datasets, as yet the database does not cover much of Africa, including Tanzania.
High resolution DEM collation
While it may be argued that elevation itself is not an environmental variable relevant to plant distributions but more as a determinant of temperature, derivations of elevation such as aspect, slope and hill-shade may be seen to play a role in delineating the spatial distributions of plants due to their effect on a given plants’ ability to photosynthesise and root successfully. Though a global DEM exists at 30′ resolution (1 km) in the form of the GTOPO30 database (USGS 1999), this may be unsuitable for calculation of shade, aspect and slope in areas of high topographic heterogeneity (Schmidt & Dikau, 1999).
As such, the use of high resolution, local scale DEM derivatives is recommended if these variables are to be included. Due to the potential disjunction between high and low resolution DEM derivatives, were high resolution local DEM derivatives be used in high resolution species envelope modelling, it is unclear whether extrapolation of high resolution local models to the regional or continental scale based on derivatives of larger scale, but lower resolution datasets such as GTOPO30 is invalid.
In an effort to provide high resolution DEMs for mountain specific high resolution Tanzanian tree climate modelling, and to verify the discontinuity between high resolution elevational derivatives and the lower resolution GTOPO30 DEM, available Tanzanian DEMs were located, collated and analysed at York. DEMs covering the Uluguru mountains, Udzungwa mountains, Chome, Mazumbai and Kihansi areas have been collected thus far. Elevational comparison of the Udzungwa high resolution DEM with the GTOPTO30 DEM (> 9000 data points) gives a spearman’s rank rs of 0.90. However, aspect, slope and hill-shade derivatives exhibit very low correlations (rs = 0.29, 0.12, 0.14 respectively). Comparison of the Uluguru high resolution DEM (> 1600 data points) shows a similar trend. Correlation of elevation gives an rs of 0.94. However, aspect, slope and hill-shade exhibit poor correlations of 0.40, 0.33, 0.18 respectively.
This indicates that discontinuities between high and low resolution DEM derivatives are so great as to prevent extrapolation of mountain scale climate models which incorporate derivatives of elevation to the regional or continental scale.
Climate surface data used to initially model plant species data
To maintain a set of climate variables relevant to a broad geographic and taxonomic range, those variables initially used are necessarily general whilst maintaining relevance for plant species. Choice was therefore based on the 8 variables proposed by Box (1981) who derived climate variables of global applicability. In addition to these, 3 further variables lacking from those suggested by Box were included to better represent temperature extremes and drought conditions, and to take advantage of the more advanced derivatives available today (Table 2).
TABLE 2. Climate Variables
*Box (1981) Variables
Average precipitation of the warmest month was not included however, due to uncertainties over it’s efficacy as a predictor, especially in equatorial regions with highly consistent intra-annual temperatures The climate surfaces are derived from the 0.05 degree Africa continental scale raster climate surfaces produced by The Centre for Resource and Environment Studies (CRES) of the Australian National University. Constructed from more than 1500 temperature stations and over 6000 rainfall stations distributed throughout Africa, this dataset has previously been tested for accuracy and interpolation errors via comparison with original Madagascan Met station data (Lees, unpublished). In order to maintain consistency with the low resolution plant species data, a separate dataset was also constructed where the climate data were resampled to an identical resolution and extent. The climate at each 1 degree plant species grid cell was calculated from the median climate value of all 0.05 degree climate cells found within. Though progress in identifying those derivatives of climate important to terrestrial organisms in temperate zones has been made by other researchers (eg. Prentice, et al. 1992), these data are specific to latitudinal ranges, and are species specific. As such they are unsuitable for investigation of continental Africa scale changes of groups of species consisting of many different functional types.
Future climate data is taken from the Hadley Centre’s third generation coupled ocean-atmosphere General Circulation Model (GCM), HadCM3B1. This scenario describes the most conservative estimate of climate change as described by the IPCC SRES report (IPCC, 2000). Future climate change scenarios were developed around three time periods of 2025, 2055 & 2085 based on 30-year averages. As the data used represents a 30-year average, projections of future species distributions do not incorporate the potential effects of increased climate variability; notably increased frequency of tropical storms and monsoon (Easterling et al, 2000; Kutzbach & Liu, 1997). The GCM data is provided at a global scale in a non-regular 2.5 x 3.75 degree grid structure. The GCM data were converted to a georeferenced lattice of points from which continuous surfaces were interpolated via an inverse distance weighting interpolation using the 8 nearest neighbours and a distance decay exponent of 2. The degree of climate change per time step (including present day GCM predicted climate for 2025) was then derived, from which the 0.05 degree CRES climate data were incremented. This methodology is universally accepted as a means to create high resolution future climate data-sets (IPCC-TGCIA, 1999; Huntley et al. 1995; Peterson et al. 2002). Future climate variables matching those used for the present climate modelling phase were then constructed from these datasets using identical methodology.
Initial publication of methodology
A paper describing the GA methodology, which describes future projections of 20 African plant species taken from the WORLDMAP database to the year 2055 was submitted to the journal: Computers, Environment and Urban Systems. The paper describes the GA methodology using a subset of the low resolution WORLDMAP plant species dataset. While demonstrating the GA methodology, the paper also concludes that due to a combination of lack of suitable conditions for the plants investigated in other areas of sub-Saharan Africa, decreasing habitat continuity under climate change and decreasing maintenance of original habitat between present distributions and those projected for year 2055, in terms of future extinction risk, narrower ranged species stand a lower chance of surviving future climate change. The extirpation estimates on which this trend was based are highly conservative, as species in the system were assumed to be free to disperse to any climatically suitable position within Sub-Saharan Africa (Tokumine et al. Submitted). The GA methods are also described in the report by David Lees.
Continental scale, large dataset richness modelling
The effects of climate change on large groups of flora at the country or continent scale are for the most part unknown despite advances in large scale modelling of the effect of climate change on certain faunas at the country scale (Peterson et al. 2002). To address this, the WORLDMAP low resolution analysis was automated and expanded, via a software program developed specifically for the task, to assess the effect of climate change on all species in the WORLDMAP database. The species envelopes were then projected to multiple timesteps, creating predictions of the effect of climate change on continental sub-Saharan African plant species richness for the years: 2025, 2055 & 2085. Patterns of species richness (SR) at each time step can often hide the flux of extirpations and colonisations which occur at each location (Mckinney & Drake, 1998). To understand the spatial pattern of these shifts, species turnovers were calculated at each time step as a measure of total species movements within a location, and as a measure of extirpations and colonisations (positive and negative turnover). The turnover projections were calculated by first calculating maps of summed extirpations (SE) and summed colonisation (SC) per time step. Total species movement was then calculated as: (SE + SC) / SR. Extirpations and colonisations were calculated as SE / SR and SC / SR respectively.
Initial projections indicate huge losses of both species and pattern of species richness distribution in the first step from the present day to 2025 (>1600 species lost), with relatively small losses of species occurring after that. Furthermore, the methodology predicts drastic changes in the distribution of species richness, especially in Zimbabwe, central Congo, and Angola (Figure 1).
Although the now-2025 timestep jump is the largest (50 years – due to data unavailability), to ensure this massive loss and change in pattern in the first timestep was not due to disjunctions between real climate and future GCM data, comparisons of present actual and present modelled data were carried out. One potential source of error identified comes from the assumption that the pattern of the GCM control run representing present day is directly representative of the pattern exhibited by the actual present day albeit at a lower resolution. If this were not the case, the delta change methodology can result in the addition of locations of difference between the control and future GCM to areas of actual present climate which differ greatly from the control GCM climate. An example of such differences can be seen in a comparison of January actual and control GCM temperatures and rainfall (Table 3).
TABLE 3: A comparison of January actual and control GCM temperatures and rainfall
The large magnitude of difference between reality and the GCM controls prompted the switch to the use of GCM only climate data for the low resolution climate modelling. It also indicates that further work needs to be carried out to determine the significance of this issue prior to finer scale modelling, for which raw GCM data is unsuitable.
The low resolution WORLDMAP analysis was therefore re-run using purely GCM data. These results indicate that although the number of extirpations remains relatively high (>1400 in the first time step), the disjunction in species richness pattern between present day and 2025 is reduced (Figure 2). A summary of turnover and species richness totals indicates gradually reducing species richness (~1/3 lost at 2025, 50% by the end of the century), with the majority of species gradually losing range size (Table 4).
TABLE 4: Summary of species turnover statistics
This range size reduction is observed across all range size quartiles. Encouragingly, for the vast majority of those species which do persist, their current range and projected future ranges overlap, indicating gradual migration may be possible within the short timescales involved. Even by 2085, ~50% of those species remaining are projected contiguous with their present day distributions.
This work is currently being prepared in a paper addressing shifts in species richness, endemism and turnover for sub-Saharan Africa.
Plan for continuing work on the continental large scale species richness projections
In order for us to make predictions for African flora in general, biases in the WORLDMAP dataset need to be addressed so that they are at least documented. At best, the dataset finally presented could be a subsample of the current WORLDMAP database, broadly representative of the whole flora. Whilst this may be impossible geographically, this is possible to broadly assess bias based on taxonomy. Work on the African plant dataset is continuing in collaboration with the Biomap Project in Bonn. Subsequent to this, endemism and species extirpations need to be mapped per timestep and species by species movements need to be better understood by tracking shifts of individual species, perhaps between WWF Biomes. This will allow us to comment on whole movements of species within areas such as the Congo basin. In addition to this, a set of HadCM3 30 year average results centred on 1995 instead of 1975 will be requested from CRU to better represent the present day control run, as these are currently unavailable and would standardise the timesteps. Once the Computers Environment and Urban systems paper is published, GA methodology, algorithms and data used will be made available on the web.
Plan of continuing work for the high resolution Tanzanian tree modelling
Prior to initiating high resolution modelling, the potential problems identified in the delta change methodology need to be addressed via further investigation and contact with the original GCM modellers at CRU. Other methods of overcoming this problem include the derivation of high resolution future climate surfaces from regional scale GCMs such as MM5, broad increases in temperature and rainfall taken from IPCC imposed across the suite of CRES variables, or a switch from the CRES data to lower resolution baseline data used for the Hadley GCMs, such as the 0.5 degree Global CRU dataset.
Report on bioclimatic modelling workshop,
Friday 26 April 2002
The aim of the meeting was to have an informal discussion where a small group of people interested in bioclimatic modelling could get together, exchange ideas and work up collaboration. Discussion was driven by a series of short presentations.
Colin McClean chaired the meeting
Si Tokumine – Genetic algorithm modelling
David Lees – Bioclimatic modelling in Madagascar
Chris Thomas – Metapopulations in bioclimatic modelling
Paul Williams – Application of WORLDMAP to bioclimatic modelling
Walter Jetz – Geometric constraints and the distribution of African birds
The meeting went very well with intelligent and useful formal discussion around the presentations, and lively informal discussion at coffee and lunch. Aside from the general networking opportunities, we agreed to investigate:
1. Merging metapopulation theory with bioclimatic modelling. This will take a practical landscape level approach to the response of organisms to changing climate i.e. the habitat patches need to be big enough and frequent enough for organisms to migrate through climate change.
2. Apply the new approach to Madagascar. This will help design of a system of corridors to mitigate habitat loss and climate change.
We aimed to get a mix of people, with out and out modellers, policy and field people. The group dynamic was excellent – we planned to finish the meeting at 15:00, but the meeting actually broke up at 18:00.
Chris Thomas, University of Leeds, firstname.lastname@example.org
Alison Cameron, University of Leeds
Mark Reed, University of Leeds, email@example.com
Paul Williams, NHM, P.Williams@nhm.ac.uk
Lisa Manne, NHM, firstname.lastname@example.org
David Lees, NHM, email@example.com
Neil Burgess, WWF, firstname.lastname@example.org
Colin McClean, University of York, email@example.com
Si Tokumine, University of York, firstname.lastname@example.org
James Taplin, University of York, email@example.com
Jon Lovett, University of York, firstname.lastname@example.org
Claire Quinn, University of York, email@example.com
Mette Termansen, University of York, firstname.lastname@example.org
Walter Jetz,Uni. of Oxford, email@example.com
Jana Schulz, University of Oxford
Box, E.O. (1981). Predicting physiognomic vegetation types with climate variables. Vegetatio, 45, 127-139.
Brookes, C.J. (2001). A genetic algorithm for designing optimal patch configuration in GIS. International Journal of Geographical Information Science, 15(6), 539-559.
Bürger, A.-M.C. (in press). A study – using WORLDMAP – of distributions of African savanna plants. In: E. Robbrecht, J. Degreef, & I. Friis (eds.). Plant systematics and phytogeography for the understanding of African biodiversity. Proceedings of the XVIth AETFAT Conference.
Busby, J.R. (1986). A biogeographical analysis of Nothofagus cunninghamii (Hook.) Oerst. In south-eastern Australia. Austral. J. Ecol, 11, 1-7.
C.A. Reynolds, T. J. Jackson, and W.J. Rawls. 1999. Estimated Available Water Content from the FAO Soil Map of the World, Global Soil Profile Databases, Pedo-transfer Functions. Data from the USDA Agricultural Research Service. Published by the NOAA National Geophysical Data Center, Boulder, CO. http://www.ngdc.noaa.gov/seg/fliers/se-2006.shtml
C.E.L.P. (2001). WORLDMAP Plant Mapping.Centre for Ecology, Law & Policy, University of York, UK.
Clifford, H.T. and Stephenson, W. (1975). An introduction to numerical classification. Academic Press, New York.
Davis, A.J., Jenkinson, L.S., Lawton, J.H., Shorrocks, B., Wood, S. (1998). Making mistakes when predicting shifts in species range in response to global warming. Nature, 391, 783-786.
De Jong, K.E., Spears, W.M. (1989). Using Genetic Algorithms to Solve NP-Complete Problems. Proceedings of the Third International Conference on Genetic Algorithms, 124-132.
Easterling, D.R., Meehl, G.A., Parmesan, C., Changnon, S.A., Karl, T.R., Mearns, L.O. (2000). Climate Extremes: Observations, modelling, and Impacts. Science, 289, 2068-2074.
Eeley, H.A.C., Lawes, M.J., Piper, S.E. (1999). The influence of climate change on the distribution of indigenous forest in KwaZulu-Natal, South Africa. Journal of Biogeography, 26(3), 595-617.
Etterson, J. R., Shaw, R. G. (2001). Constraint to adaptive evolution in response to global warming. Science. 294, 151-154.
Gordon, C., Cooper, C., Senior, C., Banks, H., Gregory, J., Johns, T., Mitchell, J., Wood R. (2000). The simulation of SST,sea-ice extents and ocean heat transport in a version of the Hadley Centre coupled model without flux adjustments. Climate Dynamics, 16, 147-168.
Graham, R.W., Grimm, E.C. (1990). Effects of global climate change on the patterns of terrestrial biological communities. Trends in Ecology and Evolution, 5, 289-291.
Hannah, L., Midgley, G.F., Lovejoy, T., Bond, W.J., Bush, M., Lovett, J.C., Scott, D., & Woodward, F.I. (2002). Conservation of Biodiversity in a Changing Climate. Conservation Biology, 16(1), 264-268.
Hill, J.K., Thomas, C.D., Huntley, B. (1999). Climate and habitat availability determine 20th century changes in a butterfly’s range margin. Proc. R. Soc. Lond. B, 266, 1197-1206.
Holland, J.H. (1975). Adaptation in Natural and Artificial Systems, The University of Michigan Press.
Hughes, L. (2000). Biological Consequences of global warming: Is the signal already apparent? Trends in Ecology and Evolution, 15(2), 56-61.
Huntley, B., Berry, P., Cramer, W., McDonald, A. (1995). Modelling present and potential future ranges of some European higher plants using climate response surfaces. Journal of Biogeography, 22, 967-1001.
Huntley, B., Webb, T.III. (1989). Migration: species’ response to climatic variations caused by changes in the earth’s orbit. Journal of Biogeography, 16, 5-19.
Intergovernmental Panel on Climate Change (IPCC). (2000). Special Report on Emissions Scenarios. Nakicenovic, Nebojsa and Swart, Rob (eds.)., Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (2001a). Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Port Chester, New York: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (2001b). Climate Change 2001: Impacts Adaptation and Vulnerability. Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Port Chester, New York: Cambridge University Press. Intergovernmental Panel on Climate Change, Task Group on Scenarios for Climate Impact Assessment.
IPCC-TGCIA, (1999) Guidelines on the Use of Scenario Data for Climate Impact and Adaptation Assessment. Version 1. Prepared by Carter, T.R., M. Hulme, and M. Lal, Intergovernmental Panel on Climate Change, Task Group on Scenarios for Climate Impact Assessment, 69pp.
J. Schmidt and R. Dikau, (1999) Extracting geomorphometric attributes and objects from digital elevation models – semantics, methods, future needs. GIS for Earth Surface Systems, 0, 154 – 171.
Johnston, K.M., Schmitz, O.J. (1997). Wildlife and climate change: assessing the sensitivity of selected species to simulated doubling of atmospheric CO2. Global Change Biology, 3, 531-544.
Kutzbach, J.E., Liu, Z. (1997). Response of the African Monsoon to Orbital Forcing and Ocean Feedbacks in the Middle Holocene. Science, 278, 440-443.
La Ferla, B., Taplin, J., Ockwell, D., Lovett, J.C. (2002). Continental scale patterns of Biodiversity: can higher taxa accurately predict African plant distributions? Botanical Journal of the Linnean Society, 138, 225-235.
Lees, D. (Unpublished). Modelling Climate Effects on Madagascar Butterflies (Phase I): Interim Report. Biogeography and Conservation Lab. Natural History Museum: London.
Lovett, J.C., Rudd, S., Taplin, J., Frimodt-Møller, C. (2000). Patterns of plant diversity in African south of the Sahara and their implications for conservation management. Biodiversity and Conservation, 9, 37-46.
Manel, S., Williams, H.C., Ormerod, S.J. (2001). Evaluating presence-absence models in ecology: the need to account for prevalence. Journal of Applied Ecology, 38, 921-931.
McKinney, M.L./ Drake, J.A., eds. 1998. Biodiversity Dynamics: Turnover of Populations, Taxa and Communities. New York: Columbia University Press.
Nakicenovic, N. & Swart, R. (Eds.) (2000). Special Report on Emissions Scenarios. Cambridge, United Kingdom: Cambridge University Press.
Nix, H.A. (1986). A biogeographic analysis of Australian Elapid snakes. In: R. Longmore (Ed.) Atlas of Australian Elapid Snakes. Australian Flora and Fauna Series, 8, (pp. 4-15).
Peterson, A.T., Ortega-Huerta, M.A., Bartley, J., Sánchez-Cordero, V., Soberón, J., Buddemeier, R.H., Stockwell, D.R.B. (2002). Future projections for Mexican faunas under global climate change scenarios. Nature, 416, 626-629.
Prentice, I.C., Cramer, W., Harrison, S.P., Leemans, R., Monserud, R.A., Solomon, A.M. (1992). A global biome model based on plant physiology and dominance, soil properties and climate. Journal of Biogeography, 19, 117-134.
Root, T.L., & Schneider, S.H. (1993). Global climate change effects on North American birds: Narrowing the mismatch between ecological studies and climate model scales. Conservation Biology, 7(2), 256-270.
Sokal, R.R., & Sneath, P.H.A. (1973). Numerical Taxonomy: The Principles and Practice of Numerical Classification. San Francisco, USA: Freeman.
Stockwell, D & Peters, D. (1999). The GARP modelling system: problems and solutions to automated spatial prediction. International Journal of Geographical Information Science, 13(2): 143-158.
Story, M., & Congalton, R.G. (1986) Accuracy Assessment: A User’s Perspective. American Society for Photogrammetry and Remote Sensing, 52(3), 397-399.
Tokumine, S., McClean, C. J., & Lovett, J. (submitted). Genetic algorithm based climate modelling of continental scale sub-Saharan African plant distributions in a GIS. Computers, Environment and Urban Systems.
USGS (1999) GTOPO30 Documentation. WWW document, http://edcdaac.usgs.gov/gtopo30/gtopo30.html.
Walter, H. (1979). Vegetation of the Earth. New York: Springer-Verlag.
Webb, T. (1987). The appearance and disappearance of major vegetational assemblages: long-term vegetational dynamics in eastern North America. Vegetatio, 69, 177-187.
Williams, P. (2000). Worldmap IV. Biogeography and Conservation Lab, The Natural History Museum, London.
Yee, T.W., Mitchell, N.D. (1991). Generalised additive models in plant ecology. Journal of Vegetation Science, 2, 587-602.