Florian Hartig Theoretical Ecology & Ecological Data Science

← Publications

Stratified aboveground forest biomass estimation by remote sensing data

Hooman Latifi, Fabian E. Fassnacht, Florian Hartig, Christian Berger, Jaime Hernández, Patricio Corvalán, Barbara Koch

International Journal of Applied Earth Observation and Geoinformation, 38, 229–241 (2015)
Cite this
@article{latifi2015stratified,
  author = {Hooman Latifi and Fabian E. Fassnacht and Florian Hartig and Christian Berger and Jaime Hernández and Patricio Corvalán and Barbara Koch},
  title = {Stratified aboveground forest biomass estimation by remote sensing data},
  journal = {International Journal of Applied Earth Observation and Geoinformation},
  volume = {38},
  pages = {229–241},
  year = {2015},
  doi = {10.1016/j.jag.2015.01.016},
}

DOI: 10.1016/j.jag.2015.01.016
Cited by 119 (Google Scholar) · 77 (OpenAlex), as of 07 September 2026

View article (DOI)

Abstract

Remote sensing-assisted estimates of aboveground forest biomass are essential for modeling carbon budgets. It has been suggested that estimates can be improved by building species- or strata-specific biomass models, but few studies have attempted a systematic analysis of the benefits of such stratification, especially in combination with other factors such as sensor type, statistical prediction method and sampling design of the reference inventory data. We addressed this topic by analyzing the impact of stratifying forest data into three classes (broadleaved, coniferous and mixed forest), comparing predictive accuracy between the strata and to a case without stratification for a set of pre-selected predictors from airborne LiDAR and hyperspectral data obtained in a managed mixed forest site in southwestern Germany. We used five commonly applied algorithms for biomass predictions on bootstrapped subsamples of the data to obtain cross-validated RMSE and r-squared diagnostics, analysed in a factorial design by ANOVA to rank the relative importance of each factor, and used selected models for wall-to-wall mapping of biomass estimates and their associated uncertainty. The results revealed marginal advantages for the strata-specific prediction models over the unstratified ones, which were more obvious on the wall-to-wall mapped area-based predictions, though further tests are necessary to establish the generality of these results. Input data type and statistical prediction method are concluded to remain the two most crucial factors for the quality of remote sensing-assisted biomass models.

What the paper shows and why it matters (AI-generated)

Splitting forest data into species strata before fitting a remote-sensing biomass model is often suggested as an accuracy booster, but few studies had systematically tested whether it actually helps once other factors are controlled for. Comparing five prediction algorithms across LiDAR and hyperspectral data with and without stratification, the authors find only a marginal gain from stratifying — input data type and the choice of statistical method matter far more — a finding that keeps getting weighed against newer stratification and biomass studies a decade on.