PhD course: Introduction to model-based inference for forest ecosystem inventories and assessments, 3 ECTS
Forest inventories traditionally rely on field sampling and design-based statistical inference. Remote sensing enables cost-efficient assessment, however, most applications of remotely sensed data for forest inventory and mapping rely on model-based rather than design-based statistical inference.
Start date: 31 March 2027 00:00
End date: 16 April 2027 00:00
Language: English
Last day of registration: 19 February 2027
Organiser: Department of Forest Resource Management
Location: Umeå
This PhD course covers basic principles of model-based inference and differences between design-based and model-based inference, with a focus on forest inven¬tories that use remotely sensed data. Assessments at the level of both individual population units and larger areas will be addressed, as well as related uncertainty measures, such as bias, variance, and mean square error of estimators and predictors. Further, it will be discussed how design-based and model-based inference can be properly combined – but also how the two frameworks are sometimes improperly mixed up. Case examples involving different combinations of remotely sensed and field data will be used during exercises.
Programme
The first part of the course will be distance-based, with video meetings and individual studies. The second part of the course will be held in Umeå, with lectures and exercises at SLU’s campus.
Course schedule outline
March 31– April 9: Distance-based lectures and individual studies
April 12 – April 16: Lectures and exercises at SLU’s campus in Umeå
A detailed schedule will be sent to participants beginning of March, 2027.
Note regarding NOVA funding
Decisions about what courses will obtain funding from NOVA are expected to be made in December 2026. In case this course would not receive NOVA funding, it will still be given, but participants would not be able to get part of their costs covered by NOVA funds.