Modelling of EIS spectrum drift from instrumental temperatures

TL;DR

This paper presents an empirical neural network-based model that accurately predicts and compensates for spectrum drift in EIS data caused by temperature variations and spacecraft motion, improving velocity measurement accuracy.

Contribution

It introduces a robust neural network model linking instrumental temperatures to spectral drift, applicable to any EIS spectrum, enhancing data correction methods.

Findings

Reproduces spectrum drift with 4.4 km/s rms error

Identifies spectral curvature and spatial offsets for correction

Applicable to all EIS spectra regardless of observing field

Abstract

An empirical model has been developed to reproduce the drift of the spectrum recorded by EIS on board Hinode using instrumental temperatures and relative motion of the spacecraft. The EIS spectrum shows an artificial drift in wavelength dimension in sync with the revolution of the spacecraft, which is caused by temperature variations inside the spectrometer. The drift amounts to 70 km s$^{-1}$ in Doppler velocity and introduces difficulties in velocity measurements. An artificial neural network is incorporated to establish a relationship between the instrumental temperatures and the spectral drift. This empirical model reproduces observed spectrum shift with an rms error of 4.4 km s$^{-1}$. This procedure is robust and applicable to any spectrum obtained with EIS, regardless of of the observing field. In addition, spectral curvatures and spatial offset in the North - South direction are determined to compensate for instrumental effects.