4.6 Article

CanopyShotNoise - An individual-based tree canopy modelling framework for projecting remote-sensing data and ecological sensitivity analysis

Journal

INTERNATIONAL JOURNAL OF REMOTE SENSING
Volume 42, Issue 18, Pages 6837-6865

Publisher

TAYLOR & FRANCIS LTD
DOI: 10.1080/01431161.2021.1944695

Keywords

-

Funding

  1. Academy of Finland under the UNITE flagship ecosystem [295100, 327211]
  2. Academy of Finland (AKA) [327211, 327211] Funding Source: Academy of Finland (AKA)

Ask authors/readers for more resources

Spatially explicit tree models for projecting remote-sensing data are limited. The CanopyShotNoise model focuses on individual-tree canopies and their interaction, growth, birth, and death processes. In simulations, superorganism formation decreased intertree distances and increased canopy gap size, while crown plasticity increased regularity of tree locations. Interestingly, in scenario C1S1, crown plasticity enhanced the effect of superorganism formation.
Very few spatially explicit tree models have so far been constructed with a view to project remote-sensing data directly. To fill this gap, we introduced the prototype of the CanopyShotNoise model, an individual-based model specifically designed for projecting airborne laser scanning (ALS) data. Given the nature of ALS data, the model focuses on the dynamics of individual-tree canopies in forest ecosystems, that is, spatial tree interaction and resulting growth, birth and death processes. In this study, CanopyShotNoise was used to analyse the long-term effects of the processes crown plasticity (C) and superorganism formation (S) on spatial tree canopy patterns that are likely to play an important role in ongoing climate change. We designed a replicated computer experiment involving the four scenarios C0S0, C1S0, C0S1 and C1S1 where 0 and 1 imply that the preceding process was switched off and on, respectively. We hypothesized that C and S are antagonistic processes, specifically that C would lead to increasing regularity of tree locations and S would result in clustering. Our simulation results confirmed that in the long run intertree distances decreased and canopy gap size increased when superorganisms were encouraged to form. At the same time, the overlap and packing of tree crowns increased. The long-term effect of crown plasticity increased the regularity of tree locations; however, this effect was much weaker than that of superorganism formation. As a result, gap patterns remained more or less unaffected by crown plasticity. In scenario C1S1, both processes interestingly interacted in such a way that crown plasticity even increased the effect of superorganism formation. Our simulation results are likely to prove helpful in recognizing patterns of facilitation with ongoing climate change.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.6
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

Article Geosciences, Multidisciplinary

Testing global and local dependence of point patterns on covariates in parametric models

Mari Myllymaki, Mikko Kuronen, Tomas Mrkvicka

Summary: A Monte Carlo version of testing for a covariate effect in a parametric point process model is proposed, with two different test statistics suggested for local investigation of the covariate effect in the globally fitted model. Simulation scheme resembling the permutation inference for GLMs is used to obtain the null distribution of the statistics, followed by a Monte Carlo test with graphical interpretation to determine the global significance of the covariate and the spatially significant areas. The proposed statistics are shown to be valuable for model construction through application to simulated and real point pattern data.

SPATIAL STATISTICS (2021)

Article Forestry

Economic losses in carbon forestry due to errors in inventory data

Roope Ruotsalainen, Timo Pukkala, Annika Kangas, Mari Myllymaki, Petteri Packalen

Summary: The study found that increasing carbon prices and reducing error levels led to decreased losses in NPV. Inclusion of carbon payments in maximizing NPV reduced the impact of errors on losses, indicating that the value of collecting more accurate forest inventory data may decrease as carbon prices rise.

CANADIAN JOURNAL OF FOREST RESEARCH (2021)

Article Biodiversity Conservation

Dispersal ability, trophic position and body size mediate species turnover processes: Insights from a multi-taxa and multi-scale approach

Soyeon Bae, Lea Heidrich, Shaun R. Levick, Martin M. Gossner, Sebastian Seibold, Wolfgang W. Weisser, Paul Magdon, Alla Serebryanyk, Claus Baessler, Deborah Schafer, Ernst-Detlef Schulze, Inken Doerfler, Jorg Mueller, Kirsten Jung, Marco Heurich, Markus Fischer, Nicolas Roth, Peter Schall, Steffen Boch, Stephan Woellauer, Swen C. Renner, Joerg Mueller

Summary: Despite increasing interest in beta-diversity, the mechanisms underlying species turnover at different spatial scales are not fully understood. Our study revealed that environmental factors and landscape spatial structure have different effects on community composition, and these effects vary among different functional groups.

DIVERSITY AND DISTRIBUTIONS (2021)

Article Mathematical & Computational Biology

Point process models for sweat gland activation observed with noise

Mikko Kuronen, Mari Myllymaki, Adam Loavenbruck, Aila Sarkka

Summary: This article aims to construct spatial models for sweat gland activation in healthy subjects and those with peripheral neuropathy using videos of sweating. Two point process models and inference methods are proposed, with the acknowledgement of potential errors in identifying sweat gland locations in the data. Image analysis steps and robust estimation procedures are necessary to address these errors.

STATISTICS IN MEDICINE (2021)

Article Environmental Sciences

Hierarchical log Gaussian Cox process for regeneration in uneven-aged forests

Mikko Kuronen, Aila Sarkka, Matti Vihola, Mari Myllymaki

Summary: A hierarchical log Gaussian Cox process (LGCP) is proposed to study the influence of one set of points on another set in point patterns, where each point in the influencing set has a parameterized signal forming an influence field. Parameters of the model are estimated in a Bayesian framework using Markov chain Monte Carlo, with Laplace approximation for the Gaussian field. The model is applied to analyze the impact of large trees on regeneration success in uneven-aged forest stands in Finland.

ENVIRONMENTAL AND ECOLOGICAL STATISTICS (2022)

Article Mathematical & Computational Biology

New methods for multiple testing in permutation inference for the general linear model

Tomas Mrkvicka, Mari Myllymaki, Mikko Kuronen, Naveen Naidu Narisetty

Summary: Permutation methods are commonly used to test the significance of regressors in GLMs for functional data sets, and new multiple testing methods have been proposed to improve power and robustness. The methods rely on sorting the permuted functional test statistics based on pointwise rank measures.

STATISTICS IN MEDICINE (2022)

Article Agriculture, Multidisciplinary

Land-use intensity and landscape structure drive the acoustic composition of grasslands

Sandra Mueller, Martin M. Gossner, Caterina Penone, Kirsten Jung, Swen C. Renner, Almo Farina, Lisa Anhaeuser, Manfred Ayasse, Steffen Boch, Falk Haensel, Janine Heitzmann, Christoph Kleinn, Paul Magdon, David J. Perovic, Nadia Pieretti, Taylor Shaw, Juliane Steckel, Marco Tschapka, Juliane Vogt, Catrin Westphal, Michael Scherer-Lorenzen

Summary: Understanding drivers and monitoring changes of biodiversity is crucial for evidence-based management and policy recommendations. Ecoacoustic monitoring offers the potential for resource-efficient ecosystem monitoring. Acoustic diversity has been shown to correlate with species richness and vegetation and landscape structure. This study found that land-use intensity and landscape structure affect species richness and composition of birds and orthopterans, indirectly impacting acoustic diversity and composition.

AGRICULTURE ECOSYSTEMS & ENVIRONMENT (2022)

Article Computer Science, Artificial Intelligence

Individual tree crown delineation in high-resolution remote sensing images based on U-Net

Maximilian Freudenberg, Paul Magdon, Nils Noelke

Summary: We propose a deep learning-based framework for individual tree crown delineation in aerial and satellite images. The method creates irregular polygons for tree crown delineation instead of bounding boxes and provides a tree cover mask for areas that are not separable. It is trainable with low amounts of training data and does not require 3D height information. The results show that the approach can efficiently delineate individual tree crowns in high-resolution optical images.

NEURAL COMPUTING & APPLICATIONS (2022)

Article Forestry

Circular or square plots in ALS-based forest inventories-does it matter?

Petteri Packalen, Jacob Strunk, Matti Maltamo, Mari Myllymaki

Summary: In airborne laser scanning (ALS)-based forest inventories, there is often a discrepancy between the circular plot shape used for model fitting and the square shape of population elements used for predictions. This study found that for equal area square and circular plots, there was no evidence of systematic prediction error when a model fitted to one shape was used to predict for the other shape. However, using a model fitted to circular plots to predict for square plots slightly underestimated the root mean square error (RMSE) value.

FORESTRY (2022)

Article Ecology

High-resolution 3D forest structure explains ecomorphological trait variation in assemblages of saproxylic beetles

Lukas Drag, Ryan C. Burner, Jorg G. Stephan, Tone Birkemoe, Inken Doerfler, Martin M. Gossner, Paul Magdon, Otso Ovaskainen, Maria Potterf, Peter Schall, Tord Snall, Anne Sverdrup-Thygeson, Wolfgang Weisser, Joerg Mueller

Summary: Climate, topography, and forest structure play important roles in shaping local species communities, but little is known about how environmental characteristics affect the functional traits of wood-living beetles involved in wood recycling. In this study, we used ecological and morphological traits of saproxylic beetles and airborne laser scanning data to investigate the factors driving the distributions of over 230 species in European temperate forests. We found that elevation and the proportion of conifers were important factors determining species occurrences, while habitat heterogeneity and forest complexity had less influence. Ecological traits, such as canopy niche, wood decay niche, and host preference, were found to be more important in shaping species responses to forest structure and environmental variation, while morphological traits showed little association with environmental characteristics. These findings highlight the potential impacts of climate and tree species composition changes on saproxylic beetle communities, and the importance of ecological traits in predicting species responses to future environmental change.

FUNCTIONAL ECOLOGY (2023)

Article Biodiversity Conservation

Analysing gap dynamics in forest canopies with landscape metrics based on multi-temporal airborne laser scanning surveys - A pilot study

Niklas Hagemann, Paul Magdon, Sebastian Schnell, Arne Pommerening

Summary: This study analyzed canopy gap dynamics in the Krycklan catchment area in northern Sweden using the Boolean model and landscape metrics. The results showed no significant trend of harmful development in forest resources between 2006 and 2015, with evidence of gaps decreasing, becoming more random, and canopy cover increasing.

ECOLOGICAL INDICATORS (2022)

Article Environmental Sciences

Predicting plant biomass and species richness in temperate grasslands across regions, time, and land management with remote sensing and deep learning

Javier Muro, Anja Linstaedter, Paul Magdon, Stephan Woellauer, Florian A. Maenner, Lisa-Maricia Schwarz, Gohar Ghazaryan, Johannes Schultz, Zbynek Malenovsky, Olena Dubovyk

Summary: This study successfully predicted above-ground biomass and species richness in grassland using remote sensing sensors and machine learning methods. The feed-forward deep neural network (DNN) outperformed the random forest (RF) algorithm in terms of prediction accuracy and generalizability.

REMOTE SENSING OF ENVIRONMENT (2022)

Article Environmental Sciences

Recreating structurally realistic tree maps with airborne laser scanning and ground measurements

J. Kostensalo, L. Mehtatalo, S. Tuominen, P. Packalen, M. Myllymaki

Summary: This study developed a framework for building structurally representative tree maps using airborne laser scanning data and ground measurements, which can accurately map the attributes of forests. Compared to other methods, this approach improves the accuracy of mapping forest attributes.

REMOTE SENSING OF ENVIRONMENT (2023)

Article Remote Sensing

Estimation of forest stand characteristics using individual tree detection, stochastic geometry and a sequential spatial point process model

Lauri Mehtatalo, Adil Yazigi, Kasper Kansanen, Petteri Packalen, Timo Lahivaara, Matti Maltamo, Mari Myllymaki, Antti Penttinen

Summary: This study discusses methods for estimating tree-level characteristics using a Horvitz-Thompson-like estimator and a new method based on a sequential spatial point process model. The new method showed comparable performance to the HT-like estimator, with lower computational demands, making it attractive for practical use.

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION (2022)

No Data Available