03.27.13: LPV analysis of exp24 with half a day output

Here, I tried to see if the noise in dQL/dt is due to lack of temporal resolution in the analysis. I compare in Figs. 1 and 2 the analysis using either 1-day and 1/2-day output. There is almost no difference suggesting that the problem is not the lack of resolution in the analysis.

Figure 1: LPV analysis for days 2015-2115.

Figure 2: Same as in Fig. 1 except that we use half-a-day output.

---

Fig. 1 was produced with statistically_stationary_test.m and Figs. 2 and 3 with plot_LPV_analysis.m using Matlab files produced by theory_test_several_cycles_script.m. All are in RESEARCH/MODELISATION/HIM/studies/diss_train_of_eddies/exp2/exp24-with-half-a-day-output/analysis_1d on the main disk on ipu1.