I, here, look at how the steady balance in the LPV equation converges in `exp2_w6` and `exp2_w6_sponge` (both have KH = 10 m2/s and MAG=500; see *this note*).

Not only the balance has not converged yet even after using 17 cycles but I do not even get a very nice PV balance between all terms using the `light` version of the LPV analysis. It would be nice to see if I recover the LPV balance using the `full` version of the analysis.

Another comment is that I use segments translated by 50 days. What I should have done before the calculation is looking at how the wave field varies. Does it have a 600-day periodicity as in `exp25` (see *this note*) or something else? I just did that and the answer is that there is **no obvious periodicity**. Time series of 1-day U between day 2000 and 3000 looks like that from a realistic eddy field; a dominant period of about 100 day but no obvious periodicity. By curiosity, I looked also at the wave field in `exp2_l`, a simulation that has a weaker wave field but zero dissipation. According to *this note* or *this note*, the LPV balance converges quite rapidly toward the steady balance (in this case, zero equals zero). Not surprising, **the wave field is actually perfectly periodic in exp2_l**, suggesting that it should be more difficult (that is, more cycles are needed) to get the steady balance in any non-periodic case such as in `exp2_w6` and `exp2_w6_sponge`.

Computed with `theory_test_light_several_cycles_script.m` and `LPV_steady_balance_test_conv.m` in `RESEARCH/MODELISATION/HIM/studies/PV_and_dissipation/diss_train_of_eddies/exp2/exp2_w6/analysis_1d` and `RESEARCH/MODELISATION/HIM/studies/PV_and_dissipation/diss_train_of_eddies/exp2/exp2_w6_sponge/analysis_1d` on the main disk on `ipu1`. The Matlab files are `diag_VC_17_cycles_100day_long_50day_apart_day2015to2815_exp2_w6_light.mat` and `diag_VC_17_cycles_100day_long_50day_apart_day2015to2815_exp2_w6_sponge_light.mat` in that same directory.