Geomagnetic Activity and GPS Anomalies at High Latitudes Adam - PowerPoint PPT Presentation

Geomagnetic Activity and GPS Anomalies at High Latitudes Adam Jacobs Mentor: Rob Steenburgh NOAA SWPC

OUTLINE Motivation Motivation Background Data Details • Alaska’s citizens use GPS for many daily activities Data Processing • Solar and Geomagnetic activity effect the accuracy of GPS Results/Plots Conclusions • GPS customers wonder when they can and can’t rely on GPS Future Work positioning (some situations Comments or are life threatening) Questions • Is there a strong correlation or predictable relationship? • If so--we want to eventually be able to forecast GPS disruption (timing/magnitude) given certain levels of geomagnetic activity Jacobs· Steenburgh Colorado REU Presentation · slide 2

OUTLINE Background Info Motivation Background Scintillation Cycle Slips Positioning Error Data Details Data Processing Results/Plots Conclusions Future Work Comments or Questions Jacobs· Steenburgh Colorado REU Presentation · slide 3

OUTLINE Scintillation Motivation Background • Solar wind and CMEs cause Scintillation perturbations in the Earth’s magnetic field Cycle Slips Positioning Error • These perturbations in turn cause particle precipitation and complex Data Details structure to develop in the Data ionosphere—especially at high Processing latitudes (top pic) Results/Plots • Large gradients in TEC—bubbles and Conclusions blobs—disrupt GPS signals (bottom Future Work pic) Comments or • Irregularities cause diffraction and Questions refraction of the GPS signal which in turn cause phase shifts and amplitude variability respectively • Scintillation refers to observed rapid Jacobs· Steenburgh changes amplitude and phase Colorado REU Presentation · slide 4

OUTLINE GPS Positioning Errors Motivation Background Scintillation Cycle Slips • The perturbed ionosphere can Positioning cause unpredictable variations Error in Total Electron Content Data Details Data • These variations can cause Processing complex signal group (pseudorange) delay Results/Plots Conclusions • Pseudorange—the apparent Future Work distance between the reference station and a satellite Questions using the time for the signal to get there • Triangulation of pseudoranges are used to get the distance Jacobs· Steenburgh Colorado REU Presentation · slide 5

OUTLINE Motivation GPS Cycle Slips Background Scintillation Cycle Slips • Occur as a result of phase shifts Positioning Error • When the phase shift exceeds Data Details the bandwidth of a receiver’s phase lock loop (PLL) the Data receiver loses lock Processing Results/Plots • The duration of loss of lock Conclusions determines the number of cycles that will slip by Future Work unrecorded(hence cycle-slips) Comments or Questions Numerous cycle slips in the GPS observations • There are several relatively create jumps in the time series. successful correction methods (Banville et. al. 2010—Radio Science) for cycle slips including polynomial fitting Jacobs· Steenburgh Colorado REU Presentation · slide 6

OUTLINE Motivation Data Details Background Data Details GPS data Geomagnetic data Data Processing Results/Plots Conclusions Future Work Questions Jacobs· Steenburgh Colorado REU Presentation · slide 7

OUTLINE Motivation GPS Data Used Background Data Details GPS data • Chose eight Alaska GPS Geomagnetic reference stations data • Various latitudes to Data check for latitude Processing dependence in GPS errors Results/Plots Conclusions • Stations are part of the Nation Geodetic Survey Future Work (NGS) Continuously Operating Reference Questions Stations (CORS) network Jacobs· Steenburgh Colorado REU Presentation · slide 8

OUTLINE Geomagnetic Data Used Motivation Background Data Details • Used daily A-index data and 3- GPS data hourly K-index data from College, AK Geomagnetic data • K-index—represents the maximum fluctuations in the Data horizontal mag. Field Processing components relative to a quiet Results/Plots day, during a three hour interval Conclusions • A-index—daily average of the eight “equivalent three hourly” Future Work (a-index) values which have Questions been converted nonlinearly from the K-index values Jacobs· Steenburgh Colorado REU Presentation · slide 9

OUTLINE Data Processing Motivation Background • Started with the daily number of cycle slips from all eight GPS stations and the A-index at College for 2003 and 2008 (quiet and Data Details active years) Data Processing • Examined the temporal and spatial relationships between the two Results/Plots data sets and calculated correlation coefficients Conclusions Future Work • Moved to GPS positioning errors (daily max and median) and Questions compared these data sets to the A-index • Created histograms of position errors • Explored higher granularity data (Fairbanks only), specifically College K-index and three hourly medians for GPS positioning errors • Examined relationships, correlations, cross-correlations, distributions, and seasonal variability (looked at specific events) Jacobs· Steenburgh Colorado REU Presentation · slide 10

OUTLINE Motivation Background Data Details Data Processing Results/Plots Cycle Slips Results/Plots Position Error Conclusions Future Work Questions Jacobs· Steenburgh Colorado REU Presentation · slide 11

OUTLINE Fairbanks Cycle Slip Correlation Motivation Background Data Details Data Processing Results/Plots Cycle Slips Daily Corr. 1 Daily Corr. 2 Daily Corr. 3 Position Error Conclusions Future Work Questions •Daily cycle slip count and College A-index data •Note excellent correlation •Also note the lack of separation between 2003 and 2008 Jacobs· Steenburgh Colorado REU Presentation · slide 12

OUTLINE PUO1 Cycle Slip Correlation Motivation Background Data Details Data Processing Results/Plots Cycle Slips Daily Corr. 1 Daily Corr. 2 Daily Corr. 3 Position Error Conclusions Future Work Questions •Strong Correlation •Excellent separation in years Jacobs· Steenburgh Colorado REU Presentation · slide 13

OUTLINE TSEA Cycle Slip Correlation Motivation Background Data Details Data Processing Results/Plots Cycle Slips Daily Corr. 1 Daily Corr. 2 Daily Corr. 3 Position Error Conclusions Future Work •Worst case out of the eight stations Questions •This station is located at a lower latitude than stronger cases Jacobs· Steenburgh Colorado REU Presentation · slide 14

OUTLINE Fairbanks 2003 Time Series Motivation Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 2008 Zoom Fair Corr. 1 Jacobs· Steenburgh Fair Corr. 2 Colorado REU Presentation · slide 15 Fair Corr.

OUTLINE Fairbanks 2003 Zoom In Motivation Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 •There doesn’t seem to be an obvious direct 2008 or lagged correlation in this data. Zoom Fair Corr. 1 Jacobs· Steenburgh Fair Corr. 2 Colorado REU Presentation · slide 16 Fair Corr.

OUTLINE Fairbanks 2008 Time Series Motivation Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 2008 Zoom Fair Corr. 1 Jacobs· Steenburgh Fair Corr. 2 Colorado REU Presentation · slide 17 Fair Corr.

OUTLINE Fairbanks 2008 Zoom In Motivation Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 2008 •Again there are no obvious correlations in Zoom 2008 Fair Corr. 1 •Interesting ~5-day periodic structure to GPS errors though Jacobs· Steenburgh Fair Corr. 2 Colorado REU Presentation · slide 18 Fair Corr.

OUTLINE Fairbanks Daily Max Error Correlation Motivation Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 2008 Zoom Fair Corr. 1 Jacobs· Steenburgh Fair Corr. 2 Colorado REU Presentation · slide 19 Fair Corr.

Fairbanks Daily Median Error OUTLINE Motivation Correlation Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 2008 Zoom Fair Corr. 1 Fair Corr. Jacobs· Steenburgh 2 Colorado REU Presentation · slide 20 Fair Corr.

OUTLINE Fairbanks 3-hourly Median Error Motivation Correlation Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 2008 Zoom Fair Corr. 1 Fair Corr. Jacobs· Steenburgh 2 Colorado REU Presentation · slide 21 Fair Corr.

OUTLINE Motivation Fairbanks Histograms Background Data Details Data Processing Results/Plots Cycle Slips Position Error Fair TS 2003 2003 Zoom Fair TS 2008 2008 • We were looking for a “circle of trust” Zoom and outer ring to subtract out noise Fair Corr. 1 Fair Corr. 2 Jacobs· Steenburgh Fair Corr. 3 Colorado REU Presentation · slide 22