Experiment on change the ion density interpolation algorithm in IPE_Plasma_Class.F90 from linear to log

[ZhuxiaoLi66]

For the log interpolation you have to change the code in only tow places. In IPE_Plasma_Class.F90 SUBROUTINE Cross_Flux_Tube_Transport

In the Q-interpolation change after !GHGM - Below the bottom of the tube at the Northern end

Change density(1:n_conv_spec,lpx,mpx) calculation to density(1:n_conv_spec,lpx,mpx) = log10(plasma % ion_densities_old(1:n_conv_spec,isouth,lp_t0(lpx),mp_t0(mpx)))i_comp_weight(1) + log10(plasma % ion_densities_old(1:n_conv_spec,inorth,lp_t0(lpx),mp_t0(mpx)))i_comp_weight(2)

after the mp,lp interpolation is done i.e. e_temperature(lpx,mpx)lp_comp_weight(lpx)mp_comp_weight(mpx)

            ENDDO
          ENDDO!
                  "add"
            ion_densities_int(1:n_conv_spec)  = 10**ion_densities_int(1:n_conv_spec)

[ZhuxiaoLi66]

3 storm test runs have been carried out after the above code change.
20230424 storm (Apr.20-25), run crashed totally at 20230425_210000, got NAN values since 20230424_03000. the number of 'Conv issue in O+ H+' in fcst files of run dir is 131703 .vs. 19724 in the control run. 20230324 storm (Mar.20-25), not crash. the number of 'Conv issue in O+ H+' in fcst files of run dir is 113461 .vs. 1211 in the control run. 20031120 storm (Nov.18-23 2003), run not crash. the number of 'Conv issue in O+ H+' in fcst* files of run dir is 168911 .vs. 29519 in the control run.

[ZhuxiaoLi66]

[ZhuxiaoLi66]