Welcome to LZN's Blog!
Wind extinguishes a candle but energizes fire.-
CESM Fully-coupled Model State Correction
Introduction
A new NCC paper provided a technology to rectify the fully-coupled model systematic bias.
The procedures are following:
- A free run (
FREE) after spin-up, of course, full of systematic biases. - A SST restoring run (
SSTR) toward observational SST.- The tech support to restore SST may refer to my previous notes (in Chinese)
- Relevant scripts for preprocessing.
- Get downward heat flux differences (SW+LW+LH+SH) from
SSTR-FREE, make it as a constant forcing and add the forcing to free run, namelyHEATR_CTRL. - Following 3, Do what you need (
HEATR_SEN).
Heat Flux Correction
I will then show the procedures to conduct the
HEATR_CTRLrun.First we lock the definition part for the cam variables to hold data input to & output from the coupler. In
$CESMROOT/models/atm/cam/src/control:! ! Public data types ! public cam_out_t ! Data from atmosphere public cam_in_t ! Merged surface data !--------------------------------------------------------------------------- ! This is the data that is sent from the atmosphere to the surface models !--------------------------------------------------------------------------- type cam_out_t integer :: lchnk ! chunk index integer :: ncol ! number of columns in chunk real(r8) :: tbot(pcols) ! bot level temperature real(r8) :: zbot(pcols) ! bot level height above surface real(r8) :: ubot(pcols) ! bot level u wind real(r8) :: vbot(pcols) ! bot level v wind real(r8) :: qbot(pcols,pcnst) ! bot level specific humidity real(r8) :: pbot(pcols) ! bot level pressure real(r8) :: rho(pcols) ! bot level density real(r8) :: netsw(pcols) ! real(r8) :: flwds(pcols) ! real(r8) :: precsc(pcols) ! real(r8) :: precsl(pcols) ! real(r8) :: precc(pcols) ! real(r8) :: precl(pcols) ! real(r8) :: soll(pcols) ! real(r8) :: sols(pcols) ! real(r8) :: solld(pcols) ! real(r8) :: solsd(pcols) ! real(r8) :: thbot(pcols) ! real(r8) :: co2prog(pcols) ! prognostic co2 real(r8) :: co2diag(pcols) ! diagnostic co2 real(r8) :: psl(pcols) real(r8) :: bcphiwet(pcols) ! wet deposition of hydrophilic black carbon real(r8) :: bcphidry(pcols) ! dry deposition of hydrophilic black carbon real(r8) :: bcphodry(pcols) ! dry deposition of hydrophobic black carbon real(r8) :: ocphiwet(pcols) ! wet deposition of hydrophilic organic carbon real(r8) :: ocphidry(pcols) ! dry deposition of hydrophilic organic carbon real(r8) :: ocphodry(pcols) ! dry deposition of hydrophobic organic carbon real(r8) :: dstwet1(pcols) ! wet deposition of dust (bin1) real(r8) :: dstdry1(pcols) ! dry deposition of dust (bin1) real(r8) :: dstwet2(pcols) ! wet deposition of dust (bin2) real(r8) :: dstdry2(pcols) ! dry deposition of dust (bin2) real(r8) :: dstwet3(pcols) ! wet deposition of dust (bin3) real(r8) :: dstdry3(pcols) ! dry deposition of dust (bin3) real(r8) :: dstwet4(pcols) ! wet deposition of dust (bin4) real(r8) :: dstdry4(pcols) ! dry deposition of dust (bin4) end type cam_out_t !--------------------------------------------------------------------------- ! This is the merged state of sea-ice, land and ocean surface parameterizations !--------------------------------------------------------------------------- type cam_in_t integer :: lchnk ! chunk index integer :: ncol ! number of active columns real(r8) :: asdir(pcols) ! albedo: shortwave, direct real(r8) :: asdif(pcols) ! albedo: shortwave, diffuse real(r8) :: aldir(pcols) ! albedo: longwave, direct real(r8) :: aldif(pcols) ! albedo: longwave, diffuse real(r8) :: lwup(pcols) ! longwave up radiative flux real(r8) :: lhf(pcols) ! latent heat flux real(r8) :: shf(pcols) ! sensible heat flux real(r8) :: wsx(pcols) ! surface u-stress (N) real(r8) :: wsy(pcols) ! surface v-stress (N) real(r8) :: tref(pcols) ! ref height surface air temp real(r8) :: qref(pcols) ! ref height specific humidity real(r8) :: u10(pcols) ! 10m wind speed real(r8) :: ts(pcols) ! merged surface temp real(r8) :: sst(pcols) ! sea surface temp real(r8) :: snowhland(pcols) ! snow depth (liquid water equivalent) over land real(r8) :: snowhice(pcols) ! snow depth over ice real(r8) :: fco2_lnd(pcols) ! co2 flux from lnd real(r8) :: fco2_ocn(pcols) ! co2 flux from ocn real(r8) :: fdms(pcols) ! dms flux real(r8) :: landfrac(pcols) ! land area fraction real(r8) :: icefrac(pcols) ! sea-ice areal fraction real(r8) :: ocnfrac(pcols) ! ocean areal fraction real(r8), pointer, dimension(:) :: ram1 !aerodynamical resistance (s/m) (pcols) real(r8), pointer, dimension(:) :: fv !friction velocity (m/s) (pcols) real(r8), pointer, dimension(:) :: soilw !volumetric soil water (m3/m3) real(r8) :: cflx(pcols,pcnst) ! constituent flux (evap) real(r8) :: ustar(pcols) ! atm/ocn saved version of ustar real(r8) :: re(pcols) ! atm/ocn saved version of re real(r8) :: ssq(pcols) ! atm/ocn saved version of ssq real(r8), pointer, dimension(:,:) :: depvel ! deposition velocities end type cam_in_tHere we note that SW and LW are from coupler, and LH and SH are from coupler. Then we will grep the relationship between these things and the variables in the output file of CAM model.
Name in code Name in output cam_out%netsw FSNS cam_out%flwds FLDS cam_in%lhf LHFLX cam_in%shf SHFLX Next, in
physpkg.F90(still this guy!), we found that lhf is calculated by cam_in%cflx (moisture flux from the surface). As the moisture flux is a constituent flux, I decide to remain the latent heat at first.Therefore, the question is to add a constant forcing to netsw, flwds, and shflx.
Luckily, we have tried to change the shflx in previous trial.
For netsw and flwds,
physpkg.F90call the radiation, and in radiation_tend module, the netsw is set. Note that inradiation.F90:real(r8), parameter :: cgs2mks = 1.e-3_r8 flds(i) = cam_out%flwds(i)*cgs2mks cam_out%flwds(i) = cam_out%flwds(i)*cgs2mksI don’t know why they organize the data structure like this. But netsw and flwds both can be handled similarly after physpkg call the radiation module.
The modified code files have been uploaded to github.
Updated 2018-11-01
- A free run (
-
CESM测试Aqua Planet试验
开始测试Aqua-Planet试验，直接按照aqua的设置run没有任何问题，但是发现SST没有annual cycle，询问师姐：
你是用的cesm的default的 aqua compset吗 如果是的话，那么好像是不需要额外读入sst文件的，而是直接在code里指定需要哪种sst profile就可以，具体是通过修改ocn_comp.F90 中 sst_option，这里不同的sst_option代表不同的profile；关于topo，好像你是需要在user_nl_cam里指定一个新的topo文件的(把高度啥的设为0)
这么说还需要自己写一个topo为0的文件。
修改轨道参数，固定黄赤交角为0度：
rb_mode='fixed_parameters' orb_obliq=0.0 orb_eccen=0.0167239 orb_mvelp=102.035Updated 2018-05-14
-
20181011交易笔记-跟得上的旋律，踩不中的鼓点
我们看下八月份的策略总结：
- 油价回调到65以下开始加仓。
- 低买offshore高卖onshore硬通货。
- 佛系加仓贵金属。
- 再次强调几条底线：50%家庭负债率红线；6个月流动性；30%风投红线。
- 少看盘多看书，多写代码搭系统。
过去两个月，1没机会，一路突突突。2没流动性，3佛了几把，4风投红线边缘试探。5已经尽力做到。
老规矩，重要的还是记录错误。过去两个月最大问题就如同题目所说，跟得上旋律，踩不中鼓点。
EWH put最高浮盈130%，但是持有到Expire被一波反弹干到-20%；随后单周VIX call上场，out of money后没有rolling，然后呢，我们看下这周：
日了狗。踩节奏难，不算清楚option price in了多少预期，out of money还没发达到套保目的。想想portfolio本身还是没什么问题，短期回撤也能接受，还是谨慎点，别上来就干，多读书了解。
宏观形势上tarrif在9月底进一步加码，纵深发展无法避免。就像Dalio所说，单纯的tarrif不是什么问题，问题是CN US两大国之间的关系在未来走向何方。时老师14年的书里讲清楚了，波动和趋势，趋势的力量怕根本不是什么川总的花边新闻，驴象扯皮能改变的。
US端不确定性是中期，以及关系上进一步财长宣布汇操。
市场情绪面上整体非常悲观。
我呢？依然乐观。最近看吴晓波《十年二十人》采访红杉沈南鹏，开始反思自己关注社会经济的时候忽略了什么。每一代人里都有佼佼者，踏踏实实做事的人，创造价值的人，慧眼寻找潜在价值并给予机会的人，减熵，为世界增加有序对而不是尸位素餐的人。他们的存在确保了社会全要素生产率水位不断提高。 前提是渠道的畅通。环境改变，渠道会彻底在正反馈下封死么？我想不会的，人性不变，但全要素生产率的水位一直向上走，每一代人的平均受教育程度也是向上的，大道之行，只是波动磕碰难免。
天佑中华，做好自己的事情，言尽于此。
Updated 2018-10-12
- A bit of experience and bitter reflection from my visit to CD
- 货币的价值
- 2018年2月5日“波动性末日”（Volmageddon）的回顾与思考
- 24年10月市场观察
- EGU23-10723 Supplementary Materials: Mangkhut (2018)
- Intel MPI specified compiler
- git submodules
- PYROMS Installation
- Coursera Machine Learning : W9 Anomaly Detection
- PGI21 WRF-CMAQ 2WAY