Table Of Contents

This Page

04.11.2013: LPV analysis of exp24 and exp34 with different resolutions

Here, I try to see if the noise that appears in dQL/dt is due to the spatial resolution. First, I have performed exp34 which is exactly like exp24 except that the resolution is doubled in x and y. I compare the LPV analysis for a small region (Figs. 1 and 2). We see that the noise (upper left panel) has a smaller resolution and is weaker but it is still present. Then, I performed the analysis for exp24 using a resolution finer than the resolution of the model grid (which is 1/10°). Doubling the resolution reproduces the same noise as for exp34 (Figs. 2 and 3) suggesting that it is totally due to a lack of resolution of the analysis, not of the model (notice that it is interesting that the results are the same with higher resolution despite the fact that the friction coefficient has not been changed between the two simulations; this suggests that the scale at which friction occurs is larger than the model resolution). With a quadrupled resolution, the noise has a smaller resolution and a noiseless pattern starts to appear (Fig. 4). Consequently, with a finer resolution, the signal-to-noise ratio increases and this may be a solution to the noise problem.


Figure 1: LPV analysis for exp24 with same resolution as the model grid (1/10°). This is the average over two cycles, days 2015-2115 and days 2065-2165.


Figure 2: LPV analysis for exp34 with same resolution as the model grid (1/20°).


Figure 3: LPV analysis for exp24 with a resolution doubled that of the model grid (1/20°).


Figure 4: LPV analysis for exp24 with a resolution quadruple that of the model grid (1/40°).

The figures were plot_LPV_analysis.m using Matlab files produced by theory_test_several_cycles_script.m. These files are in RESEARCH/MODELISATION/HIM/studies/PV_and_dissipation/forced_damped_wave/exp24/ and RESEARCH/MODELISATION/HIM/studies/PV_and_dissipation/forced_damped_wave/exp34/ on the main disk on ipu1.