Skip to main content
SLU publication database (SLUpub)

Abstract

We applied data assimilation for updating map information about growing stock volume in forests in a study area in northern Sweden, across a study period ranging from 2010 to 2022. Novel features of the study were that (i) we applied newly developed regression techniques for predicting forest characteristics from remotely sensed data, designed to avoid the bias trends that arise from standard regression methods, (ii) we applied growth models that utilized information about site quality and age, assessed wall-to-wall for the study area, based on data from repeated airborne laser scanning surveys, and (iii) we used a fully empirical approach to computing weights for the DA filter. The results showed that the accuracy obtained for predictions of growing stock volume from an initial laser scanning survey could be improved upon across the study period when a sequence of predictions using optical satellite data, digital aerial photos, and a final laser scanning survey were assimilated. Using only a sequence of optical satellite data and digital aerial photos, the accuracy of the initial laser scanning-based predictions could be maintained across the study period. The new regression methods performed better than standard regression methods in terms of avoiding bias trends, but the best overall results in terms of accuracy were obtained for standard regression combined with classical calibration. The study confirms findings from previous similar studies that data assimilation has a potential to maintain or slightly improve the accuracy of growing stock volume predictions from an initial high-quality laser scanning survey through assimilating a series of predictions from lower-quality remotely sensed data across a relatively long period of time.Nous avons appliqu & eacute; une m & eacute;thode d'assimilation de donn & eacute;es pour mettre & agrave; jour l'information cartographique relative au volume de bois sur pied pour les for & ecirc;ts d'une zone d'& eacute;tude du nord de la Su & egrave;de, sur la p & eacute;riode 2010-2022. Les aspects novateurs de cette & eacute;tude impliquent que : (i) nous avons utilis & eacute; des techniques de r & eacute;gression r & eacute;cemment d & eacute;velopp & eacute;es, con & ccedil;ues pour & eacute;viter les biais inh & eacute;rents aux m & eacute;thodes de r & eacute;gression classiques, pour pr & eacute;dire les caract & eacute;ristiques foresti & egrave;res & agrave; partir de donn & eacute;es de t & eacute;l & eacute;d & eacute;tection ; (ii) nous avons appliqu & eacute; des mod & egrave;les de croissance int & eacute;grant la qualit & eacute; et l'& acirc;ge des stations, & eacute;valu & eacute;s sur l'ensemble de la zone d'& eacute;tude, & agrave; partir d'acquisition multiples de relev & eacute;s lidar a & eacute;roport & eacute;s ; et (iii) nous avons utilis & eacute; une approche enti & egrave;rement empirique pour le calcul des pond & eacute;rations du filtre d'assimilation de donn & eacute;es. Les r & eacute;sultats ont montr & eacute; que l'exactitude des pr & eacute;dictions de volume de bois sur pied obtenues & agrave; partir d'un relev & eacute; lidar initial pouvait & ecirc;tre am & eacute;lior & eacute;e tout au long de la p & eacute;riode d'& eacute;tude gr & acirc;ce & agrave; l'assimilation d'une s & eacute;quence de pr & eacute;dictions utilisant des donn & eacute;es satellitaires optiques, des photographies a & eacute;riennes num & eacute;riques et d'un relev & eacute; lidar final. L'utilisation d'une s & eacute;quence de donn & eacute;es satellitaires optiques et de photographies a & eacute;riennes num & eacute;riques a permis de maintenir l'exactitude des pr & eacute;dictions initiales bas & eacute;es sur le relev & eacute; lidar sur l'ensemble de la p & eacute;riode d'& eacute;tude. Les nouvelles m & eacute;thodes de r & eacute;gression ont surpass & eacute; les m & eacute;thodes de r & eacute;gression standard pour & eacute;viter des biais. Toutefois, les meilleurs r & eacute;sultats globaux en termes d'exactitude ont & eacute;t & eacute; obtenus avec la r & eacute;gression standard combin & eacute;e & agrave; un & eacute;talonnage classique. Cette & eacute;tude confirme les r & eacute;sultats d'& eacute;tudes similaires ant & eacute;rieures : l'assimilation de donn & eacute;es permet de maintenir, voire d'am & eacute;liorer l & eacute;g & egrave;rement, l'exactitude des pr & eacute;visions de l'accroissement des volumes suite & agrave; une premi & egrave;re campagne de balayage lidar de haute qualit & eacute;, gr & acirc;ce & agrave; l'assimilation d'une s & eacute;rie de pr & eacute;visions issues de donn & eacute;es de t & eacute;l & eacute;d & eacute;tection de moindre qualit & eacute; sur une p & eacute;riode relativement longue.

Keywords

Data assimilation; data fusion; forest inventory; mapping; regression analysis

Published in

Canadian Journal of Remote Sensing
2026, volume: 52, number: 1, article number: 2623684
Publisher: TAYLOR AND FRANCIS INC

SLU Authors

UKÄ Subject classification

Forest Science
Earth Observation

Publication identifier

  • DOI: https://doi.org/10.1080/07038992.2026.2623684

Permanent link to this page (URI)

https://res.slu.se/id/publ/146133