Statistics at Forest Research Geoff Morgan. Forestry Commission Formed in 1919 as a response to the shortage of timber caused by the war. Manages about.

Statistics at Forest Research

Geoff Morgan

Forestry Commission

• Formed in 1919 as a response to the shortage of timber caused by the war.

• Manages about 1m hectares of woodland (Forest Enterprise)

• Awards grants and gives out licences (conservancies)

Forest Research

• Research agency of the Forestry Commission

• Two main sites– Alice Holt, near Farnham– Northern Research Station, near Edinburgh– about 12 field stations– 280 staff

FR Divisions

• Tree Health (Pathology and Entomology)

• Ecology

• Forest Management

• Environmental and Human Sciences

• Biometrics, Surveys and Statistics

Some key problems• Plant diseases (e.g., Phytophthora ramorum)

• Insect pests (e.g., Hylobius)

• Natural regeneration/direct seeding

• Grey squirrels and deer

• Monitoring and surveying

• Landscape ecology

• Bio-diversity

• Climate change

Pine Weevil

Hylobius abietis

• Adult females lay eggs in bark of stumps

• Adults emerge 1-2 years later

• Attack newly planted trees

• Can weevil be controlled without using pesticides?

The problem

Hylobius experiments

• Size of eggs

• Shallow/deep feeding

• Time to flight

• How to estimate weevil numbers

Seedling survival

• Is natural regeneration of Corsican Pine possible?

• Are sufficient viable seeds produced?

• What conditions lead to successful germination?

• Do the seedlings survive?

Seed Predation by mice

The experiments

• Do mice have preference for different seed types?

• Do different conditions discourage predation?

• Use of GLM/GLMMs

• More general models via MCMC approach

Growth of young trees

• Experiments/studies measure tree height and diameter at regular intervals

• Examining the effects of – weeding treatments– vegetation– pruning

• Use of end point analysis repeated measures or growthcurves?

Some statistical problems

• How many observations?

• Plot sizes - how many trees is optimal?

• Blocks - shapes and sizes

• Random effects/error strata

• Multivariate approaches

Working with scientists An evolutionary relationship

• Software gets easier– Excel– GenStat

• Statistics gets harder

• Servant to team member

• Quality Assurance

In summary

• Diverse

• Frustrating

• Challenging

• All in a wonderful setting!