This changelog records the changes in the 3D rockfall trajectory model Rockyfor3d since the beginning of development.
Changes in version 6.1 (April 2026)
-
After an extensive optimisation and verification process using Leave-One-Out Cross-Validation (LOOCV), we concluded that the most robust and realistic version of Rockyfor3D is one without the Ff and Sf function introduced in December 2024. The core of Rockyfor3D will now remain unchanged for the years to come. This optimisation process was based on 377 mapped fresh deposited rockfall blocks (with volumes ranging from 0.05 to 450 m3) from rockfall and rock mass fall events at 38 different sites in Austria, Canada, France, Greece, Italy, Slovenia, Spain, Sweden, Switzerland and the United States. We thank all ecorisQ members that have been willing to share their data. Detailed information on the data and the optimisation process will follow in a scientific paper, in which all contributors will be co-author.
Changes in version 6.0.2 (December 2025)
-
Adapted the scaling function Sf = 1 / ((Vn1/15.24)^2 + 1); This results in less conservative tangential energy loss during the rebounds and therefore much less optimistic trajectories than in v6.0.1
Changes in version 6.0 / 6.0.1 (December 2024)
- Corrected occurrence of NaN values in the calculation screen (net output) .txt files (v6.0.1)
- Adapted the Logfile to the v6.0
- Correction of cell corner problems that led to excessive velocities/energies
- Reintroduction of the friction and scaling function in the energy loss (rebound) calculation during ground impacts
- Introduction of the total kinetic energy correction based on the total apparent kinematic coefficient of restitution proposed by Noël et al. (2023)
- Introduction of the maximum rotation velocity based on Caviezel et al. (2021)
- Added the option “cliff area weighted” number of simulated trajectories from each start cell to account for the difference in productivity of start cells with different slope angles
- Added the option rapid automatic simulation with low and medium roughness
- Added new options in the command line version
- Corrected errors in the tree impact detection
- Adapted the function that controls energy loss due to the impact height on the tree stem
- Corrected errors in the net impact detection
Changes in version 5.2 (March 2016)
- Adapted the class values of the energy and passing height output rasters
- Added the propagation probability output raster (called Propag_probability.asc) for providing the area specific spatial occurence probability for risk calculations
Changes in version 5.2 (2014)
- Revision of the Graphical User Interface – adding user-friendly simulation options
Changes in version 5.1 (2014)
- Added E_50.asc, E_90.asc, E_95.asc, E_98.asc, E_99.asc and Ph_50.asc, Ph_90.asc, Ph_95.asc, Ph_98.asc, Ph_99.asc as output rasters to provide different percentiles of total kinetic energy and passing height per cell
Changes in version 5.0 (2011)
- Translation of the Matlab code into C/C++
- Adapted the Rn value for asphalt roads
- Adapted the rolling/small rebounds condition for blocks moving upslope
- treefile.txt needs to be provided without a header
Version 4 (2007 - 2011)
- First stand-alone version of Rockyfor3D based on Matlab (using the MCRInstaller)
- Formalisation of the calculation of the reach probability in Rockyfor3D v4 by L. Dorren (in 2010) using: number of passages * 100 / (number of simulations per source cell * number of source cells) [in %]
Version 3 (2004 - 2007)
- First Matlab-based version using 3D vectors called RockyFor published by Dorren et al., (2006)
Version 2 (2001 - 2003)
- First simplified, Matlab-based version of Rockyfor (see Dorren et al., 2004)
Version 1 (1998 - 2001)
- ROCKY3: Matlab-based predecessor versions of Rockyfor (see Dorren et al., 2003)
- Slope profile based Matlab simulations and spatially explicit variants leading to ROCKY3