MEGAN3 issues

The following issues have been identified and will be fixed for the next version (MEGAN3.1) expected to be released in late summer 2019.

1. Error in the beta version of DB_emissions.csv (input file for MEGAN_EFP). The Soil NO emission factors had the wrong emissions.

FIX: This was fixed for the "released" MEGAN3.0 version. If you are using the "beta" version then replace your DB_ emissions.csv file with the new version at the web site.

2. Fixed bug in MEGCAN.f code (routine LeafEB): Tdelt was included on line 1089 but should not be. This was not a problem for the compiler MEGAN3 was tested on (Tdelt was assigned a value of 0) but it causes a problem on others (that assign a value of NaN or something else).

FIX: On line 1089 replace

IRoutairT = LeafIR(TairK + Tdelt, Eps)


IRoutairT = LeafIR(TairK, Eps)

The original test case was replaced with one created after fixing this bug.

3. Lack of emission data for regions outside of the US. As a result, the input files that are currently provided with MEGAN3.0 will only work for locations in the contiguous US.

FIX: We will have input files that cover the rest of the world provided for MEGAN3.1 (Expected release: Late summer 2019).

4. Ecosystems with no trees are assigned a global default that is a mix of large number of tree species. This has little or no impact on results since there are no or few trees at these locations but it will result in some strange tree distributions.

FIX: We will assign "unknown tree" to these locations for any MEGAN3.1 input files using this.

5. Update Sqlite libraries. Run time error from src/ "sqlite3.OperationalError: no such column: Description Vegetation.VegID "

FIX: This may be caused by having an old version of sqlite libraries.

6. Prepmegan4cmaq issue for longitude = 0. Interpolation made by prepmegan4cmaq in points corresponding to 0° longitude get a 0 value for at least some variables.

FIX: We'll try to get the developers of this regridding code to fix this (this code was not created specifically for MEGAN). Other possibilities for regridding MEGAN input files include using ARCGIS. Note that this is only a problem if you are calculating emissions in a region where longitude = 0 (western Europe and Africa).

7. Preprocessor crashes or fails to compile. This was not a problem for the compiler we tested MEGAN3.0 on but it is a problem for other compilers

FIX: In area_mapper.f90 code, add the following two lines before “allocate(wrkid(mxetype),wrkid_cnt(mxetype),stat=astat)” (line 789).

if (allocated(wrkid)) deallocate(wrkid)

if (allocated(wrkid_cnt)) deallocate(wrkid_cnt)