FoRSense discussion group

Dear colleagues,

A new year, a new FoRSense! We would very much like to invite you to join the FoRSense discussion group meeting on the 30^th of May at 16:00 in room Lumen 7 (in-person only). For those not in the know (yet), FoRSense is a platform for presentations, discussions and connecting with fellow PhD candidates on anything forest- and remote sensing-related with a drink in hand! We have a Teams group (link) in which we will make announcements and in which you can sign up to present, organize a presentation, or discuss a topic for upcoming sessions.

For the upcoming session, Natalia Malaga Duran, a PhD candidate from the GRS group will be presenting her work on the impact of the spatial configuration of national forest inventory plots on the representation of aboveground biomass (AGB) density in biomass map units. She will be defending her PhD in September and has indicated that she would very much appreciate input on the last chapter of her thesis.

Currently the manuscript is still being finalized, but a preprint will be shared (in confidence) beforehand, so that everyone participating can prepare for the session and ensure a spicy discussion! Natalia has already provided an abstract, see below:

While National Forest Inventories (NFIs) serve as the primary data source for country-level aboveground biomass (AGB) estimates, many tropical countries still face challenges in completing and updating their inventories. Meanwhile, advancements in space-based biomass products and future satellite missions offer promising opportunities to complement ground-based information for AGB estimation, which is crucial for greenhouse gas (GHG) mitigation and adaptation efforts. Yet, integrating remote sensing-derived products with NFI information poses several harmonization challenges, including differences in the spatial support between sample units. The study aims to assess the accuracy of five NFI plot spatial configurations (one single- and four cluster plot designs) for estimating mean AGB density within fixed-size rectangle (8-16 ha) representing a biomass map unit in a tropical and temperate forest sites. Employing a discrete bottom-up modelling approach by means of a hierarchical marked point process (HMPP) framework, we incorporate the intrinsic dynamics and spatial patterns of different tree size classes and environmental conditions within forest stands in shaping AGB densities. Through 1000 simulations, we evaluate the accuracy and precision of the mean AGB density estimates of the NFI-inspired plots sampled within the large map unit.

Our results show that our HMPP models accurately represent the AGB distribution observed within the per-site remote sensing-based map units. As anticipated, higher-level point intensities influenced the distribution of smaller trees within forests, while clustering of larger trees was observed due to environmental conditions and natural disturbances. This study reveals that the spatial configuration of a plot does bare differences in accurately representing the AGB density within a specific remote sensing-based map unit, with the cluster-cross configuration emerging as the most accurate and precise across both forest sites. We also identified the sensitivity of some plot configurations, notably the clustered L-shaped configuration, to fixed spatial trends. Our study contributes to understanding the impact of plot spatial configuration on map-to-plot intercomparison analysis, essential to any application integrating ground-based information with remote sensing-derived products.

We are very much looking forward to seeing you there!

Cheers,

Eva, Sietse and Arjen