TL;DR
This paper introduces the first permanent, automated terrestrial laser scanning (TLS) measurement station designed for continuous, high-resolution monitoring of boreal forest canopy dynamics. The station provides unprecedented long-term 3D structural and phenological data, enabling new insights into forest growth, seasonal changes, and even circadian movements.
What Problem Does This Paper Address?
While Terrestrial Laser Scanning (TLS) is a powerful tool for vegetation monitoring, its potential for long-term, high-resolution studies of forest dynamics remains largely untapped. Existing research often relies on short-term campaigns or lacks the automation and continuous data streams needed to capture subtle, seasonal, or daily changes. Furthermore, readily available, long-term TLS datasets for researchers are scarce. This paper addresses this gap by presenting a novel, permanent TLS station that overcomes these limitations, providing a continuous, detailed 3D view of a boreal forest ecosystem.
Methods
The core of this work is the development and deployment of an automated TLS measurement station. Key methodological aspects include:
- Automated Data Acquisition: A RIEGL VZ-2000i TLS scanner, housed in a weather-protected hood, is programmed to perform scans every hour. This high frequency (one scan per hour) captures both short-term (circadian) and long-term (seasonal) dynamics.
- High Spatial and Temporal Resolution: The system is configured for a 0.006° angular resolution, achieving a point spacing of approximately 0.01 m at 100 m distance. This ensures detailed structural information.
- Data Management and Transmission: A robust workflow is established for data acquisition, local temporary storage, and daily overnight transfer to a central network storage (NLS-NAS). This ensures data integrity and accessibility.
- Data Pre-processing: Raw .rxp files are converted to the ASPRS .laz format, and then further processed to generate analysis-ready 3D point clouds and raster images (TIFF) for individual trees and the entire scanned area. This includes point cloud filtering and rasterization using equirectangular projection.
- Phenological and Structural Analysis: The paper demonstrates the application of the collected data for monitoring long-term phenological changes (e.g., leaf-out) and short-term circadian movements in individual trees. This involves techniques like volumetric change analysis and point cloud clustering.
Datasets
The primary dataset introduced in this paper is the continuous TLS time-series data collected from the Hyytiälä research forest in Southern Finland. This dataset spans the growing season of 2020 (April to July) and includes hourly scans, resulting in approximately 3,000 scans and 23 TB of data. The paper also leverages complementary data from the SMEAR II station, providing local weather parameters (wind speed, temperature, humidity, etc.) at a 1-minute interval, which are crucial for contextualizing the TLS measurements.
Key Insights and Ideas
- Unprecedented Temporal Resolution for Forest Dynamics: The hourly scanning frequency is a significant leap, allowing for the detection of subtle daily and seasonal changes that are missed by traditional, less frequent TLS campaigns.
- Bridging the Gap Between Short- and Long-Term Monitoring: The station effectively captures both rapid circadian movements and slower seasonal growth patterns, offering a holistic view of forest dynamics.
- The Power of Permanent, Automated Systems: The automation and permanent nature of the station are critical for generating consistent, long-term datasets, overcoming the limitations of manual, campaign-based studies.
- TLS as a Non-Destructive Tool for Phenology: The study highlights how TLS can accurately track phenological events like leaf-out, providing a valuable alternative to traditional methods.
- Potential for Circadian Rhythm Research: The ability to collect data under varying light conditions (including night) and with high temporal resolution opens new avenues for studying plant circadian rhythms in natural environments.
Key Results
- Successful Deployment and Operation: The permanent TLS measurement station was successfully installed and operated, collecting hourly scans throughout the 2020 growing season.
- Detailed Phenological Tracking: The time-series data clearly captured the start of leaf-out in a Silver birch tree, correlating with meteorological data and demonstrating the system's ability to detect seasonal phenological changes.
- Quantification of Circadian Movements: The study demonstrated the detection of circadian movements in a Silver birch tree over a 30-hour period, with displacements up to 5 cm observed in branches.
- Accurate DBH Estimation: The TLS data allowed for the automatic monitoring and estimation of Diameter at Breast Height (DBH) for individual trees, showing a consistent growth trend over time with an accuracy comparable to manual measurements.
- High-Density Data Generation: Individual scans produced approximately 600 million points, contributing to the large volume of data collected (23 TB for the season).
Limitations and Future Work
- Environmental Condition Sensitivity: The detection of circadian movements is sensitive to environmental factors like wind and precipitation. Future work aims to mitigate this by combining TLS with other sensor types and by carefully selecting data acquisition periods.
- Stem Occlusions: Understory vegetation and dense canopies can cause stem occlusions, particularly for DBH measurements. This is noted as a challenge, especially for coniferous trees, and suggests the need for more advanced segmentation and analysis techniques.
- Data Handling and Processing: The large volume of data generated requires efficient processing techniques. The authors suggest the development of new processing methods to handle this data effectively.
- Automation of DBH Estimation: While DBH estimation was demonstrated, further work is needed to fully automate this process for a wider range of tree species and conditions.
- Extension to Other Ecosystems: The authors propose extending this concept to other monitoring sites and networks, including agricultural areas, to broaden the application of TLS for vegetation dynamics monitoring.
Summary updated