The plight of the natural world is more and more becoming something of concern; with reports bombarding us about the reduction and potential loss of animals and plants in the wild. Plus the more local the loss or threat to a species is to us the more people are recognising the importance of conservation. And yet conservation projects, as with any project must be measurable, otherwise how do conservationists know whether their efforts are successful or having an impact of any kind?
The People’s Trust for Endangered Species has been at the heart of conservation in the UK and across the world since 1977. Despite its focus on raising funds and providing support for conservation projects the PTES still undertakes conservation work itself including a number of surveys. These surveys are as important as the conservation activities themselves and are an integral part of any conservation programme.
The PTES surveys and studies are run in order to help understand population numbers and changes; without understanding how a population is changing over time it is almost impossible to decide what, if any, conservation actions are required. Likewise once a project is underway the conservationists need to see whether the actions are making a difference to the species in question, and it’s here that the conservationists at PTES use statistical analysis, in the form of GenStat.
“It’s too big a task to count all the individuals in a population, so a sample is counted sometimes from sightings, sometimes from signs (tracks or droppings), which comes with its own “statistical error”, explained David Wembridge, of the PTES. “In addition, biological populations are naturally variable – population sizes fluctuate from year to year as prey-predator and environmental interactions change, and in order to spot underlying trends (whether a population is stable or declining in the long-term, for example), it’s necessary to statistical analyses of the data, in particular, so called trend analyses – which is what we use GenStat for.”
The problem with spotting underlying trends is that long datasets are needed, with data collected over a period of ten years or more. Hence the PTES run ongoing surveys, which are repeated annually; for example one such survey – the National Dormouse Monitoring Programme has been running for 22 years.
The graph below is a typical example of how the PTES use the survey data to understand the dormouse population.
Hazel dormice population index 1991-2007
“There is a significant, approximately linear, decline in the years up to 2000. This appears to continue after 2000, but levels off over the last couple of years.”
“From 2005 the population estimate exhibits oscillations of up to 25%. The 2007 population has a strong influence on the shape of the curve, so it will be interesting to see whether the 2008 results are high, suggesting the stabilisation of the population, or low, in which case the curve will be dragged back down again.” (National Dormouse Monitoring Programme, interim report 2008)
Although conservationists at the PTES use a small part of GenStat, it’s another example of how GenStat is an important facet in a bigger picture. Without the trusted and reliable statistical analyses to indicate changes to populations conservationists might struggle to decide where to deploy their efforts and actions.
Our thanks to the PTES for their help and advice with this piece. The PTES run several surveys – to find out more or to get involved look at the PTES website. For more information on other conservation projects using GenStat please go to the GenStat Conservation Pages.
The correlation matrix describes the correlation among a number of different variables. In GenStat you can view this matrix graphically using the DCORRELATION procedure. An example of the correlation within and between two datasets is given below (to replicate go to Help -> Examples -> DCORRELATION).
Using the strip underneath as a key, high positive correlations between variables are given by dark red colours, whilst high negative correlations are given by dark blue colours. Greens, yellows and light blues indicate low correlation. (Note that the diagonal of the correlation matrix is not plotted, though this can be added via the SHOW option.) A key in the upper right hand corner indicates the position of the variables in our calculations, e.g. the red square shows the correlation between N and Axis_2. The dashed black line indicates the separation of our two datasets, so anything within the black lines can be considered the correlation between the datasets; using the PLOT option we can separate the three parts of our plot, as seen below.
Here we have used the COLOURS option to specify the colours we want representing the correlation values -1, 0 and 1 (blue, white and red respectively).
The DCORRELATION procedure can also be used on a single dataset or any symmetric matrix whose values are between -1 and 1, such as similarity matrices.
Our colleagues from the University of Wageningen are running a three day course on Introduction to Mixed Model QTL Mapping Using GenStat at the Department of Genetics, ESALQ/USP, Piracicaba, SP. Avenida Pádua Dias, 11. Bairro Agronomia, Brazil on 13 to 15 December 2010. The course is designed for graduates and professionals interested in a flexible QTL mapping approach,applicable in standard situations (SIM, CIM) as well as more specialized situations (multi-environment QTL mapping, QTLxE, multi-trait QTL mapping, association mapping, in- and outbreeders). Familiarity with mixed models and quantitative genetics is recommended. For more information and to book your place please go to the website.
Our training schedule is always being updated according to user requests and requirements, so do email support with suggestions and any specific requirements you have and check our website for new courses.
Carey Biggs is attending and presenting at ASC/OZCOTS, in Fremantle Australia, between 6 and 10 December.
Simon Harding & David Baird, two of the VSN International software developers, will be presenting at the International Biometric Conference in Florianopolis, Brazil between 5 and 10 December.
Other events we are attending or supporting in 2011 include: SUSAN at the University of Botswana, Gaborone, Botswana from 27 June – 1 July 2011, the Australasian Applied Statistics Conference, Palm Cove, Tropical North Queensland from 12-15 July 2011. 58th ISI World Statistics Congress in Dublin, Ireland, from 21- 26 August 2011.
If you would like to meet with the VSNi staff at one of these events please email support to arrange the details.