ASReml-R Snippets: Constraining Variance Components

Bees & Elephants

Using mainly the GLM functionality within Genstat, large datasets are quickly and easily analysed with the results factored back into future experimental design.  The conservation of the African elephant provides tremendous opportunities for simultaneously conserving biodiversity and increasing benefits to local communities. Their role as a flagship species of global significance helps maintain biodiversity of the ecosystems they inhabit. In addition, the cultural and aesthetic values of elephants are also important, not only to African societies, but to the world at large. But this needs to be balanced against the fact that as the African elephant range becomes more and more fragmented and elephants get confined into smaller pockets of suitable habitat, humans and elephants are increasingly coming into contact and in conflict with each other. Tracking and recording the actual numbers of elephants in Africa is difficult; data are collated using a wide variety of techniques from ground or aerial surveys to dung counts and informed and other ‘guesstimation’. Overlay the obvious difficulties in the range and distances elephants can travel (global satellite tracking is helping this process) means putting accurate numbers on elephant populations can be, excuse the pun, a mammoth task. That said, counts are achieved and have shown the elephant population in Kenya has gone through various changes. In 1973, the population was about 167,000 declining to just 19,000 in 19891. The African Elephant Status Report 20072 (AESR 2007), recently published by the African Elephant Specialist Group (AfESG), puts the definite elephant count at 23,353, up 1,317 on the 2002 survey results. Add in probable, possible and speculative counts and the 2006 figure is 31,636; a rise of 2,830 on the 2002 figures. Alongside the elephant population growth, the human population in Africa has escalated way beyond anyone’s predictions, reaching an estimated 924 million3 in 2006. This increasing human population is starting to encroach dramatically into traditionally wildlife rich areas forcing large animals like elephants (and lions, buffalos and leopards incidentally) into smaller and smaller areas. Of course, 6 tonne animals like elephants have a huge requirement for food and water and naturally migrate across the landscape searching for sustenance. This is bringing them into direct conflict with people as new villages, roads, schools, fences, bridges and farms are being built over these natural wildlife corridors. The human-elephant conflict is becoming a serious issue in Kenya as people are now killing, spearing or poisoning elephants that come onto their land to eat their crops. So while elephants are major economic assets in generating tourist revenue, crop raiding by elephants is a cause of increasing levels of conflict between local people and wildlife management strategies across Africa. The economic damage to small-scale farmers can be crippling. Unless a cost-effective method of limiting crop damage is found, the pressure on wildlife managers to implement radical methods such as culling, will inevitably become a reality. As Lucy herself says, “When you meet people who are affected by these invading elephants you can only feel the deepest sympathy that they have lost their entire harvest. This causes genuine stress, damage and sometimes even death to these poor rural villagers.” #### **What can be done?** One of the methods that has been tried with variable success is the installation of electric fences to keep wild animals, particularly elephants, out of people’s farms. However, this is very expensive, requires a reliable supply of electricity and cannot be financed by the farmers so they become reliant on donors. Lucy works with a research team called Save the Elephants, which is headed by the elephant expert Dr Iain Douglas-Hamilton OBE; an organisation dedicated to developing a tolerant relationship between elephants and humans. Her research is furthering the aim to find crop-protection methods that can be financed and managed by the farmers themselves. The use of chilli peppers, either as disincentive crop or as a deterrent spray has been pioneered by Osborn (2002) in Zimbabwe with promising, albeit limited, success. Other variations include light-weight fencing hung with cow bells and laced with a potent mix of red chilli and black engine oil, thought to both alarm and irritate the elephants’ huge, sensitive trunks. However, Lucy’s research, based upon Kenya folklore which says that elephants are afraid of bees, would be an innovative and socially appropriate solution. #### **David and Goliath** The project explores the use of traditional, wooden beehives as both an elephant deterrent and a social and economic boost to poverty-stricken rural communities through the sustainable harvesting of honey. Her earlier research has shown that elephants react with obvious alarm and retreating behaviour to bee sounds. Further, recent studies show that elephants will actually avoid trees with beehives. By setting hives in acacia trees and recording how elephants react to them, Lucy hopes to be able to demonstrate that beehives can be used as an eco-barrier protecting fields of crops. When she isn’t persuading the British Army to help her fix her overworked Land Rover, Lucy works from a green canvas tent alongside the beautiful Ewaso Ng’iro river, in the Samburu Reserve.Day to day work involves setting up sample plots and selecting trees for beehives to be hung in. Many plots are tested which means driving through acres of park (hence the overworked Landie) and frequent encounters with elephants. Lucy has employed a local assistant Lukas, from the Ndorobo tribe; well known for their beekeeping abilities, and Lukas has 13 hives of his own which he helps to support his family with. There are also 4 long term elephant monitors, 1 MSc student from Nairobi University studying debarking effects of elephants and 1 other Kenyan researching whether or not lions use the same wildlife corridors as elephants. In addition, a team of researchers, all local Samburu are trained to identify all the elephants in the reserve; incredibly they know about 750 elephants by sight alone #### **Data Analysis** As well as the bee data, data are also collected from observations into factors such as temperature, wind direction and time of day to establish if the elephant behaviours observed are due to these factors, or if the elephants are indeed responding to the presence of the bees. Using mainly the GLM functionality within Genstat, large datasets are quickly and easily analysed with the results factored back into future experimental design. #### **Now, next steps and the future** In this day and age with all the issues of environment and climate change to be able to find a way of assisting farmers without destroying the natural habitat must be of the utmost importance. Here we have an opportunity to be a part of such a programme; and additionally to prove once again to the non-statistician that data analysis and statistics are a crucial part of everyday living. [http://elephantsandbees.com/](http://elephantsandbees.com/)

Birdlife International

BirdLife attempted to identify and record all present waders and coastal birds. From there they can identify the areas that need protection and preservation or even rehabilitation.  #### **What was their goal?** Their aim was to raise awareness of the decreasing populations in coastal birds and waders. They hope this will help the protection of the estuarine and tidal ecosystems around Australia. BirdLife attempted to identify and record all present waders and coastal birds. From there they can identify the areas that need protection and preservation or even rehabilitation. #### **What did they use Genstat for?** Generalised Linear Models (GLM) were used to determine if statistically significant trends in the maximum counts were occurring. Count numbers were regressed against years to detect trends. Analysis was undertaken using Genstat 13.0 software. Data was assessed for fit to assumptions of GLM (normality, spread etc). For species that were found to greatly violate these assumptions, log transformation was undertaken. Transformations of data were undertaken in Genstat.

MIKE Monitoring Elephants

It’s not news that elephants are endangered; for literally thousands of years ivory has driven hunters to kill these magnificent beasts. Combine this deliberate destruction with increased land use change and habitat transformation by private enterprise, international aid and government expansion across Africa and these magnificent beasts are still under threat even within the national parks and other protected areas set aside for them.  However, before we get too carried away with horrific idea that elephants may well become extinct; there is another side to this. Two separate taxa, or races, of elephants exist in Africa today: the savannah elephant, which is well known and has been studied and valued by the international tourist trade for decades, and the diminutive forest elephant, which until recently has remained in obscurity beneath the trees. In the Congo Basin, an area noted for being one of the great wilderness areas with the second largest rainforest on earth – it is home to an outstanding range of biological diversity; forest elephants have until recent decades thrived. These types of wildernesses have always been recognised for their high conservation value, being home to a wide variety of flora and fauna, and as such provide invaluable resources of timber, minerals and other natural resources. The Congo Basin is a typical case where the resources available are in high demand in rich, industrial nations and are also critical to the economies of poor nations. And it is at this point we in the Western developed world need to be very careful about conservation and conservation techniques; the international conservation efforts are aimed at managing and promoting sustainable natural resource exploitation and functional ecosystems. It is incredibly important that we recognise the needs of people in the developing world, and try to ensure they are able to exploit their natural resources in ways that do not fundamentally alter the ecology of the environment, upon which millions of rural people still depend. The high value of resources in these wilderness areas means that it is only fair that Africans have an opportunity to exploit and benefit from them and are encouraged to do so legally, ethically and with an eye on their own local ecology. Complications occur as industrial expansion pushes into the wildernesses to obtain these natural resources, an expansion that relies on a network of roads and rail infrastructure. It is this increase in roads that appears to be causing some of the biggest threats to the forest elephants. Illegal elephant killings tend to occur close to roads, since these areas are easily accessible, so the density of forest elephants in the protected areas is determined, not by the size of the area but by how many roads criss-cross through the wilderness. Elephants are by their nature, shy and avoid human contact as much as possible, so how are they reacting to the increasing number of roads across their home ranges? Will they adopt a “siege” mentality and reduce their home range to avoid roads, or a “skirmish” strategy, where they will continue to range widely, irrespective of the number of roads? I spoke with Stephen Blake of the Wildlife Conservation Society and Max Planck Institute of Ornithology who has recently headed up a study into this very issue. His team fitted a large sample of forest elephants in the Congo Basin with GPS telemetry collars, so that they could track their roaming behaviour. They collected data from 28 elephants located in conservation areas in the Congo, Central African Republic and Gabon. Figures 1 & 2 below show data collected from 28 elephants located in the Congo, Central African Republic and Gabon using Genstat. Figure 1: Relationship between area of MCP home range and the area of road-less polygon; y axis – Home range size (kmsq), x axis roadless polygon size (kmsq).  Figure 2: Maximum linear displacement versus area of road-less polygon; y axis maximum linear distance moved (km), x axis roadless polygon size (kmsq).  “The data we collected covered the sex of the elephant, the site (the protected area in which the elephants were collared), the number of roads in the area, the number of days the elephants were collared,” explained Stephen. “I initially analysed the result collected in Genstat, using linear regression for data exploration followed by fitting standard curves to the data (Figures 1 and 2). Genstat was vital in the research programme as I can trust the statistical procedures. I was able to talk directly with developers when I came up against any queries. I have found the Genstat support team to be extremely helpful, I usually get a response to a query within 24 hours and often much sooner. We could have gone much further with the analysis, but in order to understand the influence of roads on forest elephant ranging, the regression options in the Genstat menu showed us clear statistical patterns.” Results indicated that the forest elephants routinely crossed protected roads, however only one elephant crossed an unprotected road, and of these one did so, as far away from the nearest village as she could, and her speed increased as she crossed the road. “Our study showed that elephants seem to adopt a siege mentality with the encroaching roads”, said Stephen Blake, “and the issue with this is that over time the elephants will have less and less access to food, in quantity and variety, leading to food competition and aggression; both of which are likely to lead to a reduction in population.” “The main issue seems to be that elephants are risk averse, hence why they avoid roads where they can. So here we have a problem; central African states need the income earned from the natural resources found in their countries, but not only do we not want to see the elephants themselves disappear, a change like that to any ecosystem is likely to have more wide ranging impacts.” “There are options for policies to overcome the issues, before it is too late,” says Stephen. “Based on our research, we recommend that no new roads are built on existing forest elephant conservation areas, but to access the natural resources near these sites temporary roads should be built, so that once the resources are exhausted the area can easily return to natural wilderness. The other option is to try to reduce the elephants’ fear of roads, by preventing settlements on the roads, enforcing anti-poaching. These are of course expensive options, but without addressing this we face losing these beautiful creatures.”