Irregular Satellites of the Planets

Probing Magnetic Reconnection with Active Region Transient BrighteningsMartin DonachieAdvisors: Adam Kobelski & Roger Scott1OutlineBackgroundARTBs: What are they?An overview of typical characteristicsPreparing the dataSearching for ARTBsPlotting light curvesNumerical modelingConclusionsActive Region Transient Brightenings22ARTBs: What are they?Short-lived X-ray brightenings Caused by magnetic reconnectionFirst observed in 1983Based on data from images from SXT Japanese scientists identified a clear set of events that they called ARTBs (1992)Combining SXT data with simultaneous observations of hard X-rays and radio led to the conclusion that ARTBs are most likely scaled down flares

Active Region Transient Brightenings33PropertiesSudden brightenings of magnetic loopsEnergy released: 1025-1029 ergs (c.f. flares: ~1032 ergs)Hard X-ray fluxes ~10-3 < regular flaresFrequency: 1-40 events per hour in active regions (c.f. a few flares per day for active Sun; < 1 per week for quiet Sun)Timescale: a few, to tens of minutes

Active Region Transient Brightenings4Why study them?To better understand the mechanisms behind magnetic reconnectionOccur at lower energies than flares, BUT, far more frequentlyThey can be treated statisticallySince they occur on a smaller scale a number of simplifying assumptions can be madeEasier to describe with current numerical models

Active Region Transient Brightenings5Ultimate goal is to better understand the mechanisms behind magnetic reconnection.Looking at previous slide can see that they occur at lower energies, however5The DataXRT data from HinodeActive region 11112Aluminium-on-mesh filterStart time: 11-10-2011 0:09End time: 16-10-2011 23:48xrt_prep already runCleans up data: removes dark currents, cosmic ray hits, hot pixels etc.Original data set parsed into multiple smaller sets~60 new sets, 30 mins several hoursAverage cadence ~ 80-180 seconds

Active Region Transient Brightenings6Say why we had to split into smaller arrays: computing time etc.Wrote some code that kept running total of elapsed time, and when greater than 30 mins and a sufficient gap presented itself the data was separated.6Searching for ARTBsUse find_artb.proCalculates the running mean and standard deviation of the backgroundSearches for pixels 5 above noise levelWhen such a pixel is found:Program flags itLooks at adjacent pixels to determine whether they are connected to same eventTraces out an abstract, 3D, shape which is one ARTB regionWhen no more connected pixels are found, starts searching for new ARTB region

Active Region Transient Brightenings77Searching for ARTBsResult: new array, elements = -1, except where ARTB regions lieUse a binary count to track where each region lies: a=2n-1 (n is ARTB #)To extract information from a particular region:Enter chosen value of nDivide whole array by aARTB region lies where array = 1When two regions overlap slightly more complexActive Region Transient Brightenings88Active Region Transient Brightenings9

9Light CurvesActive Region Transient Brightenings10find_artb.pro effectively creates a mask for each ARTB regionCombining each mask with original data:Can plot light curves for each ARTBFirst: select ARTB via method on previous slideSum region along Z directionGives total area covered by chosen regionMultiply resultant area with original data along entire image stackGet X-ray data for that ARTB alone

10Light CurvesActive Region Transient Brightenings11

11Active Region Transient Brightenings12

Have to isolate light curves in tim12Next StepsImproving light curve codeCounting problemBounding problemModelingFeed information from light curves into the numerical modelActive Region Transient Brightenings13ModelingEnthalpy-based thermal evolution of loops (EBTEL) model0D: only one value of temperature, pressure and density at any given timeExcellent agreement with more sophisticated modelsUses 4 orders of magnitude less computing timeModel ARTBs as bundles of unresolved strands evolving independentlyFit multiple parameters to EBTEL model in order to match observed data

Active Region Transient Brightenings14Input ParametersPeak heating (ergs cm-3 s-1)Heat pulse width (s)Strand length (cm)Strand delay (s)The length of time between strand n firing, and strand n+1No. of strandsActive Region Transient Brightenings15Input ParametersInferred from flare loop dataScaled down to provide initial search valuesCode uses an amoeba to search parameter spaceChanges shape to find best fitSet size of steps for each parameterBigger steps: faster computation but may skip over best fit valuesSmaller steps: slower but better chance of finding fit valuesActive Region Transient Brightenings16Peak heating(ergs cm-3 s-1)Heat pulse width (s)Strand length (cm)Strand delay (s)No. strandsInput0.3401007x1085100Scale range10101x10715Best fit0.322117.8437.109x1089.614107.768Input ParametersActive Region Transient Brightenings17Example: Sub-set 14, ARTB #1Active Region Transient Brightenings18

Emission Weighted Average TemperatureActive Region Transient Brightenings19

Energy over timeActive Region Transient Brightenings20

Problems EncounteredInitial goal was to gather enough data points to treat statisticallyNot enough quality data to get good fitsMost ARTBs too short to gather sufficient data pointsNeed higher cadence data and/or longer eventsActive Region Transient Brightenings21ConclusionsDetection method is effectiveCurrent amoeba model for searching parameter space is too sensitive to initial guessesNeed a more robust way of finding best fit parametersActive Region Transient Brightenings22