ashiklom
commented
4 years ago pre 1979 data with the met data I have
Key issue there is making sure that the products have the same temporal resolution and the same variables. See the ED2 wiki on the met driver file for more details on the ED_MET_DRIVER_HEADER file format (which describes the format of the HDF5 files).
4.5 hours/year
This seems really slow. My runs were taking on the order of <5 minutes per year. I would make sure that you compiled ED2 for optimization (--kind E), not debugging. You may also want to experiment with how frequently you're writing various output files --- the run speed scales more-or-less linearly with output file frequency (2x more frequent files makes the runs roughly 2x slower). In particular, for the history files, you may want to consider writing the files only annually or monthly until the month you need, then restarting from that month with higher history file frequency.
It looks like the datasets were generated some how with PECAN tools but I have failed to be able to install the PECAN software on my local or on pic
Yes, making PEcAn easier to install has been on our to-do list for a while. In theory, you should be able to install the PEcAn.data.atmosphere package and the PEcAn.ED2 packages by downloading the PEcAn source and then, from the root directory, running make .install/modules/data.atmosphere and make .install/models/ed, respectively. But there are a lot of dependencies, including potentially some system libraries, so let me know if you have questions. The general workflow is that PEcAn.data.atmosphere::download.* functions are used to download meterology * to the PEcAn standard format, and PEcAn.ED2::met2model.ED2 is used to convert from PEcAn standard format to ED2-specific format.
depending on ED configuration SLA and Vcmax are prescribed parameters constant over time or they can be updated if ED is set up with plant trait plasticity
Just to be clear here, SLA and Vcmax are prescribed once at the beginning of the run either way --- there is no straightforward way to prescribe time series of these parameters or modify them (without stopping the run, tweaking the configs, and re-starting the run; see my notes on that issue). It's just that when trait plasticity is off, the prescribed SLA and Vcmax are the values that are always used, whereas when trait plasticity is on, the value of each parameter at any given time point is a function of light level: e.g. SLA_used[cohort] = SLA + x * light_level[cohort].
kdorheim
Progress