Forecast Geomagnetic Secular Variation via NASA Geomagnetic Ensemble - PowerPoint PPT Presentation

Forecast Geomagnetic Secular Variation via NASA Geomagnetic Ensemble Modeling System (GEMS) Weijia Kuang, NASA Goddard Space Flight Center with contributions from BW Project: bavk Andrew Tangborn (UMBC) Ce Yi (SSAI) Terence Sabaka (GSFC) Tianyuan Wang (NOAA) NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Forecast Geomagnetic Secular Variation via NASA Geomagnetic Ensemble Modeling System (GEMS) Weijia Kuang, NASA Goddard Space Flight Center 1. Geomagnetic secular variation (SV) is of fundamental importance 2. Decadal SV forecast is feasible, but is computationally challenging 3. BW project aims to find cost-effective geomagnetic data assimilation (GDAS) NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic SV affects very much our life News Sport Weather Shop Earth Travel Jersey Jersey airport runway re-named as magnetic pole shifts 15 October 2014 Jersey In addition to water and air, our life depends also on geomagnetic field! The changes should last for 56 years, the airport said Jersey's runways will be re-numbered on Wednesday night as island aviation authorities catch up with the planet's shifting magnetic field. NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

170˚ 170˚ 170˚ 170˚ 170˚ 170˚ 170˚ Geomagnetic SV affects very much our life 2019 180˚ 180˚ 180˚ 180˚ 180˚ 180˚ 180˚ Magnetic Pole shift from geomagnetic model (IGRF, WMM) News Sport Weather Shop Earth Travel 2015 − 170˚ − 170˚ − 170˚ − 170˚ − 170˚ − 170˚ − 170˚ − 160˚ − 160˚ − 160˚ − 160˚ − 160˚ − 160˚ − 160˚ 2010 Jersey − 150˚ − 150˚ − 150˚ − 150˚ − 150˚ − 150˚ − 150˚ Jersey airport runway re-named as magnetic pole shifts 80˚ 80˚ 80˚ 80˚ 80˚ 80˚ 80˚ 2005 15 October 2014 Jersey − 140˚ − 140˚ − 140˚ − 140˚ − 140˚ − 140˚ − 140˚ In addition to water and air, our life depends also on 2000 7 7 7 7 7 7 7 8 8 8 8 8 8 8 ˚ ˚ ˚ ˚ ˚ ˚ ˚ geomagnetic field! − 130˚ − 130˚ − 130˚ − 130˚ − 130˚ − 130˚ − 130˚ 7 7 7 7 7 7 7 6 6 6 6 6 6 6 ˚ ˚ ˚ ˚ ˚ ˚ ˚ 1990 1980 − 120˚ − 120˚ − 120˚ − 120˚ − 120˚ − 120˚ − 120˚ 74˚ 74˚ 74˚ 74˚ 74˚ 74˚ 74˚ 1970 1960 1948 72˚ 72˚ 72˚ 72˚ 72˚ 72˚ 72˚ 70˚ 70˚ 70˚ 70˚ 70˚ 70˚ 70˚ The changes should last for 56 years, the airport said 1904 1831 Jersey's runways will be re-numbered on Wednesday night as island aviation authorities catch up with the planet's shifting magnetic field. NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019 110˚ 110˚ 110˚ 110˚ 110˚ 110˚ 110˚ ˚ ˚ ˚ ˚ ˚ ˚ ˚ 90˚ 90˚ 90˚ 90˚ 90˚ 90˚ 90˚ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1

Geomagnetic SV holds the key information of Earth’s interior • It is a dominantly dipole field at surface • It originates from the Earth’s liquid core http://www.esa.int/spaceinimages/Images/2013/11/Earth_s_magnetic_field NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic SV holds the key information of Earth’s interior • It is a dominantly dipole field at surface • It originates from the Earth’s liquid core • It displays complex spatial and temporal variations Plots are based on the CM4 model (Sabaka et al 2004) NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic SV holds the key information of Earth’s interior • It is a dominantly dipole field at surface • It originates from the Earth’s liquid core • It displays complex spatial and temporal variations Non-dipolar magnetic field at CMB over the past 400 years from gufm1 (Jackson et al 2000) and CM4 (Sabaka et al 2004) NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic SV holds the key information of Earth’s interior Streamline of convective flow in Magnetic field line generated by core the outer core convection • It is a dominantly dipole field at surface • It originates from the Earth’s liquid core • It displays complex spatial and temporal variations • It is generated and maintained by the convection in the Earth’s fluid core (geodynamo) Geodynamo process (visualization of simulation results) NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic SV holds the key information of Earth’s interior • It is a dominantly dipole field at surface • It originates from the Earth’s liquid core • It displays complex spatial and temporal variations • It is generated and maintained by the convection in the Earth’s fluid core (geodynamo) Net magnetic energy change from kinematic -> magnetic energy transfer and Ohmic dissipation (simulation results) NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic data assimilation (GDAS) is unique for fundamental research and societal application Observed Br at CMB in 1990 Truncated simulated Br at CMB Observed Br at CMB in 1990 Simulated Br at CMB Numerical geodynamo models simply cannot reproduce observations! NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic data assimilation (GDAS) is unique for fundamental research and societal application Observed Br at CMB in 1990 Truncated simulated Br at CMB Observed Br at CMB in 1990 Simulated Br at CMB Numerical geodynamo models simply cannot reproduce observations! GDAS can help improve the models! NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

NASA GEMS: the framework for geodynamo simulation and geomagnetic forecast GDAS Driver 𝐲 " 𝐲 $ 𝐲 $ 𝐲 " EnKF Analysis System MoSST Geodynamo x f : forecast System 𝐲 " = 𝐲 $ + 𝐋 ' 𝐳 − 𝐈 ' 𝐲 $ 𝛃 : dynamo parameters y: observation 𝐐 $ ' 𝐈 , 𝜖𝐲 $ x a : analysis 𝜖𝑢 = 𝐍 𝐲 $ , 𝛃 𝐋 = 𝐈 ' 𝐐 $ ' 𝐈 , + 𝐒 𝐈 : observation operator 𝐲 $ 𝑢 " = 𝐲 " R : observation error 𝐲 $ , 𝐐 $ = 𝐲 $ − . 𝐲 $ − . 𝐲 $ covariance NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Geomagnetic SV forecast is feasible (old results)… Observed B r (GUFM1 + CM4) Comparison of geomagnetic secular variation forecasts (Kuang et al 2010) Forecasted B r from GEMS (20-year analysis cycle) NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

But GDAS is computationally very expensive Algorithm A hybrid pseudo-spectral scheme (on spherical surface) and a finite difference scheme (in radius) Estimated resolution ∆ℎ ~ 𝐹 8/: 8/= ~ 𝐹𝑆 < = 8/: ∆𝑢 ~ ∆ℎ 𝑆 < 𝑆 < = 𝐹 = 10 @A (current values) 𝑆 < ~ 10 @B , 𝐹 ~ 10 @8C (For Earth’s core) Numerical grid in meridional surface NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

But GDAS is computationally very expensive CPU expense of Geomagnetic data assimilation N ens = 1 N ens = 10 N ens = 100 N ens = 1000 “Earth-like” Current asssimilation assimilation NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

What is our BW project? Find the computationally cost-effective geomagnetic data assimilation (GDAS) approach 1. The optimal ensemble size with full covariance analysis? 2. A working hybrid covariance using small ensemble sizes? NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Optimal ensemble sizes are possible! Mean forecasted B r at CMB in 1990 10 ×( O-F ) at the CMB NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019

Summary 1. Geomagnetic secular variation (SV) is of fundamental importance CPU expense of Geomagnetic data assimilation 2. Decadal SV forecast is feasible, but N ens = 1 is computationally challenging N ens = 10 N ens = 100 3. BW project aims to find cost- N ens = 1000 effective geomagnetic data assimilation (GDAS) showed “Earth-like” possible optimal ensemble sizes assimilation Current 4. Next step: search for a working assimilation hybrid covariance for GDAS NCSA BW Symposium, Sun River, Oregon, June 2-6, 2019