Estimating forest cover in the presence of missing observations

Abstract

In the first sample-based Danish National Forest Inventory (NFI) initiated in 2002, measurements were not obtained for some plots owing to there being insufficient time to complete the measurements for the inexperienced measurement crews and lack of access to some plots on privately owned land. Nationally, the percentage of missing observations ranged from 1.8 to 7.8% in different years. Missing observations may cause the complete case estimator of forest cover mean and variance to be biased. An unbiased estimator of forest cover for NFI data with missing observations was derived, which is almost as efficient as the full data estimator. The estimator of the standard error of the estimate is unbiased, giving the correct coverage when calculating the 95% confidence interval (CI). Applied to the Danish NFI data, the estimator gave a forest cover estimate of 12.4% with an estimated standard error of 0.002576, resulting in a 95% CI of 11.9–12.9%. This corresponds to an estimated total of 534,888 ha forest land with confidence limits of 513,159–556,618 ha.

Keywords:

• Confidence interval
• estimator
• mean
• NFI
• standard error

Acknowledgements

The first author was supported by the Danish Natural Science Research Council, grant 272-06-0442.