The three approaches for applying the MEGAN model framework are described below:
1) INTEGRATED: MEGAN has been integrated directly into coupled physics/chemistry/biology 1D canopy and boundary layer models (e.g., SOSA), regional air quality models (e.g., WRF-chem) and global climate/earth system models (e.g., CESM, GEOS-chem, and E3SM). In each model time step, the MEGAN driving variables are generated and the MEGAN emissions impact atmospheric chemistry enabling direct representation of BVOC interactions and feedbacks. These models mostly use inputs (weather and landcover) from their own modeling systems and algorithms and emission factors based on MEGAN version 2 or 2.1.
2) SITE: These MEGAN codes are driven by a time series of inputs at a single site. The codes have been written in Visual Basic, Matlab and Python. They are primarily used to investigate BVOC flux and concentration observations at individual field locations, to assess the MEGAN model at these sites, and to examine the sensitivity of the MEGAN model to various model inputs and processes. These models include different versions of the MEGAN framework (e.g., version 2.1, version 3).
3) UNCOUPLED: Codes have been developed to estimate off-line MEGAN emissions for regional air quality models (e.g., CMAQ and CAMx) using WRF model weather outputs. The publicly available FORTRAN code, MEGAN2.1, is described by Guenther et al. 2012 (see https://gmd.copernicus.org/articles/5/1471/2012/) and has been widely used for regional to global modeling in the past decade. It provides the user with global maps of emission factors and landcover distributions. The MEGAN2.1 User guide, FORTRAN code, input files, global BVOC emissions output and a test case are all available at doi.org/10.5281/zenodo.10525127. The MEGAN2.1 code has also been implemented in the flexible GEE-MEGAN framework described by Zhang et al. and publicly available at https://doi.org/10.5281/zenodo.15714886. GEE-MEGAN enables integration of dynamic satellite-derived vegetation inputs using Google Earth Engine.
MEGAN3 has more transparency allowing users to view, modify and customize the emission factors, landcover and other input variables to provide domain specific inputs. This is especially useful for simulations of urban areas that cannot be accurately characterized using the MEGAN2 global data sets. This may include ultra high resolution (e.g., 0.5 to 10 m) landcover and ground surveys of landcover and urban vegetation emission factors. MEGAN3 also includes improvements to the canopy environment model, light and temperature response algorithms, emission response to canopy depth, and stress responses. A set of python preprocessor programs were written to merge growth forms, ecosystem types, species composition data and species-specific emission factors for MEGAN3 versions 3.0, 3.1 and 3.2. For MEGAN3.3, the preprocessor has been incorporated into the main FORTRAN code so that a separate python code is not required.
MEGAN3.21 is available for download from zenodo:
Code: https://doi.org/10.5281/zenodo.10526206
Inputs: https://doi.org/10.5281/zenodo.10939297
MEGAN3.3 is still being tested