Integrating Remote Sensing and Machine Learning to Assess Forest Health and Susceptibility to Pest-induced Damage
Bibliographic record
Abstract
Spruce budworm (Choristoneura fumiferana; SBW) outbreaks are cyclically occurring phenomena in the northeastern USA and neighboring Canadian provinces. These outbreaks are often of landscape level causing impaired growth and mortality of the host species namely spruce (Picea sp.) and balsam fir (Abies balsamea (L.) Mill.). Acknowledging the recent SBW outbreak in Canadian provinces like Quebec and New Brunswick neighboring the state of Maine, our study devised comprehensive techniques to assess the susceptibility of Maine forests to SBW attack. This study aims to harness the power of remote sensing data and machine learning algorithms to model and map the susceptibility of forest in terms of host species availability and abundance (basal area per hectare; BAPH, and leaf area index; LAI), their maturity and the defense mechanism prevalent. In terms of host species abundance mapping our study explores the integration of satellite remote sensing data to model BAPH and LAI of two economically vital SBW host species, red spruce (Picea rubens Sarg.) and balsam fir, in Maine USA. Combining Sentinel-1 synthetic aperture radar (SAR), Sentinel-2 multispectral, and site variables, we used Random Forest (RF) and Multi-Layer Perceptron (MLP) algorithms for modeling LAI and BAPH. The results demonstrated the superiority of RF over MLP, achieving smaller normalized root mean square error (nRMSE) by 0.01 and 0.06 for LAI and BAPH, respectively. Notably, Sentinel-2 variables, especially the red-edge spectral vegetation indices, played a significant role in both LAI and BAPH estimation, with the minor inclusion of site variables, particularly elevation. In addition, using various satellite remote sensing data such as Sentinel-1 C-band SAR, PALSAR L-band SAR and Sentinel-2 multispectral, along with site variables, the study developed large-scale SBW stand impact types and susceptibility maps for the entire state of Maine. The susceptibility of the forest was assessed based on the availability of SBW host species and their maturity. Integrating machine-learning algorithms, RF and MLP, the best model, utilizing site (elevation and aspect) and Sentinel-2 data achieved an overall accuracy of 83.4% to predict SBW host species. Furthermore, combining the host species data with age data from Land Change Monitoring, Assessment, and Projection (LCMAP) products we could produce the SBW susceptibility map based on stand impact types with an overall accuracy of 88.3%. Moreover, the work builds upon the assessment of susceptibility of SBW host species taking into account the concentration of several canopy traits using remote sensing and site data. The study focused on various foliar traits affecting insect herbivory, including nutritive such as nitrogen (N), phosphorous (P), potassium (K), and copper (Cu), non-nutritive such as iron (Fe) and calcium (Ca), and defensive parameters such as equivalent water thickness (EWT) and leaf mass per area (LMA). Using Sentinel-2 and site data, we developed trait estimation models using machine-learning algorithms like Random Forest (RF), Extreme Gradient Boosting (XGB), and Support Vector Machine (SVM). The accuracy of the developed model was evaluated based on the normalized root mean square error (nRMSE). Based on the model performances, we selected XGB algorithm to estimate Ca, EWT, Fe, and K whereas Cu, LMA, N, and P were estimated using RF algorithm. Regarding the variables used, almost all the best performing models included Sentinel-2 red-edge indices and depth to water table (DWT) as the most important variables. Ultimately, the study proposed a novel framework connecting the concentrations of foliar traits in SBW host foliage to tree susceptibility to the pest, enabling the assessment of host susceptibility on a landscape level. To sum up, this study highlights the advantages and effectiveness of integrating satellite remote sensing data for enhanced pest management, providing valuable insights into tree attributes and susceptibility to spruce budworm outbreaks in Northeast USA. The findings offer essential tools for forest stakeholders to improve management strategies and mitigate potential forthcoming SBW outbreaks in the region.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".