DOECLIM variables & time series

[bpbond]

These guys:

    std::vector<double> temp;
    std::vector<double> temp_landair;
    std::vector<double> temp_sst;
    std::vector<double> heatflux_mixed;
    std::vector<double> heatflux_interior;
    std::vector<double> heat_mixed;
    std::vector<double> heat_interior;
    std::vector<double> forcing;

really should use the Hector time series class.

On a very related note, the whole TemperatureComponent::getData structure is bizarre, with a big if/else block depending on whether a date is supplied or not. This should be restructured to be like simpleNbox.

[kdorheim]

Reorganizing how the variables are returned is an easy fix. As for migrating to the Hector time series class I am not sure what kind of list that would be. From my understanding, the std::vector calls were used because the meat of the code is an implementation of an analytical solution of DOECLIM, not the system of differential equations.

The C++ changes to the Hector time series class is probably over my head. At one point I think we even spoke about redoing it to use the system of differential equations instead of the analytical approximation but I don't really know.

[bpbond]

Let's shelve it for now.

[bpbond]

The getdata section was the big one, in that it had bad behavior.

[kdorheim]

Could this be a v3.5 milestone?

[bpbond]

Yes!