CINDI Workshop BIRA Brussels 2010 3 10 Pandora

CINDI Workshop, BIRA, Brussels, 2010 -3 -10 Pandora Direct Sun Data during CINDI Alexander Cede Nader Abuhassan Jay Herman

1. This talk 2. Short overview of Pandora 3. Pandora direct sun data during CINDI

The Pandora project - NASA saw the need for a ground-based instrument network that can be used for validation of satellite retrievals of trace gas amounts and aerosol properties.

The Pandora project - NASA saw the need for a ground-based instrument network that can be used for validation of satellite retrievals of trace gas amounts and aerosol properties. - Desirable is a network like Dobson-Brewer network for total ozone or AERONET for aerosol properties. AERONET stations

The Pandora project - NASA saw the need for a ground-based instrument network that can be used for validation of satellite retrievals of trace gas amounts and aerosol properties. - Desirable is a network like Dobson-Brewer network for total ozone or AERONET for aerosol properties. - In addition there was the idea of regional “denser” networks to study the sub-satellite-pixel variability of highly variable species (e. g. aerosols, NO 2). This would help to determine the best compromise between temporal and spatial resolution of the GEO-CAPE mission (geostationary orbit). AERONET stations

Different remote sensing techniques Satellite nadir view - needs profile information - needs albedo information - no profile information retrieved - scattered light from “larger region”

Different remote sensing techniques Satellite nadir view - needs profile information - needs albedo information - no profile information retrieved - scattered light from “larger region” Ground MAXDOAS - can retrieve some profile information - scattered light from “local region” - total column needs air mass factor calculations

Different remote sensing techniques Satellite nadir view - needs profile information - needs albedo information - no profile information retrieved - scattered light from “larger region” Ground MAXDOAS - can retrieve some profile information - scattered light from “local region” - total column needs air mass factor calculations Direct sun - no profile information needed (accurate total columns) - no profile information retrieved - small “cone” in atmosphere

The Pandora project As a consequence the Pandora project was born in 2006 with 2 main requirements:

The Pandora project As a consequence the Pandora project was born in 2006 with 2 main requirements: 1) The instrument should be able to measure both direct sun and sky spectra.

The Pandora project As a consequence the Pandora project was born in 2006 with 2 main requirements: 1) The instrument should be able to measure both direct sun and sky spectra. 2) The instrument should be designed for network operation. Therefore the total instrumental cost should be kept as low as possible.

The Pandora project As a consequence the Pandora project was born in 2006 with 2 main requirements: 1) The instrument should be able to measure both direct sun and sky spectra. 2) The instrument should be designed for network operation. Therefore the total instrumental cost should be kept as low as possible. The idea was that for calibration Pandora could be compared to stable low noise systems like MFDOAS from WSU.

Pandora is a spectrometer system that measures (ir)radiance with a narrow field of view (~1. 6° full angle) pointing at the sun, the moon, or the sky It consists of - a head sensor - a tracker - an optical fiber - a spectrometer - a temperature controller - a computer that operates the instrument and processes the data

Pandora head sensor

Pandora head sensor Fused silica window

Pandora head sensor OPEN ND 1 POL 240 ND 2 POL 120 ND 3 POL 0 ND 4 First filterwheel

Pandora head sensor OPEN OPAQUE U 340 OPEN BP 300 OPEN Second filterwheel

Pandora head sensor Fused silica plano-convex lens

Pandora head sensor Fused silica optical fiber

The tracker Use tracking system from „Directed. Perception“. Advantages: - Commercially available - Low cost (<3000$) - Comes with easy to control serial interface Disadvantages: - Gears wear out (so need maintanance, some units more, some less) - Has no optical encoders - Limited zenith motion (no downwards looking) à We already have started to design our own tracker; this has the full 4 -range of motion and optical encoders

The fiber…

The fiber Fibers have lot’s of issues … Transmission changes with bending Transmission changes at reconnection

The fiber… Fibers have lot’s of issues … but we decided to use them Transmission changes with bending Transmission changes at reconnection Service loop FC connectors

The fiber… Fibers have lot’s of issues … but we decided to use them Transmission changes with bending Transmission changes at reconnection Biggest advantage: System becomes modular! Service loop FC connectors

The fiber… Fibers have lot’s of issues … but we decided to use them Transmission changes with bending Transmission changes at reconnection Service loop FC connectors Biggest advantage: System becomes modular! “Long fiber forgets incidence location, azimuth, and polarization, but remembers incidence zenith angle” Therefore, careful with non-diffuse input!

The spectrometer… The standard Pandora wavelength range is 270 -530 nm and a resolution of ~0. 5 nm (similar OMI).

The spectrometer… The standard Pandora wavelength range is 270 -530 nm and a resolution of ~0. 5 nm (similar OMI). We tested mini-spectrometers from different providers. - Stellar. Net: many issues - Ocean. Optics: resolution has strong dependence on temperature - Hamamatsu: no custom design possible; also strongly assymetric slit function - Avantes: stable, symmetric slit function; issues with read out electronics

The spectrometer… The standard Pandora wavelength range is 270 -530 nm and a resolution of ~0. 5 nm (similar OMI). We tested mini-spectrometers from different providers. - Stellar. Net: many issues - Ocean. Optics: resolution has strong dependence on temperature - Hamamatsu: no custom design possible; also strongly assymetric slit function - Avantes: stable, symmetric slit function; issues with read out electronics à We decided to use the Avantes spectrometer But we are also testing a “hybrid” system (Avantes optical bench with Hamamatsu Detector and read out electronics)

The temperature controller Commercially available system from „TE Technology“. - Serial interface - Controls temperature at feed-back position to about ± 0. 2°.

The temperature controller Commercially available system from „TE Technology“. - Serial interface - Controls temperature at feed-back position to about ± 0. 2°. Note: Despite stabilizing the outside of the spectrometer to ± 0. 2° the electronics board temperature still typically varies for ± 1° and sometimes more

Current status of Pandora - Head sensor: very mature; gets a camera in the next version

Current status of Pandora - Head sensor: very mature; gets a camera in the next version - Temperature controller: we are trying to improve the heat sink

Current status of Pandora - Head sensor: very mature; gets a camera in the next version - Temperature controller: we are trying to improve the heat sink - Tracker: own design expected to be tested later this year

Current status of Pandora - Head sensor: very mature; gets a camera in the next version - Temperature controller: we are trying to improve the heat sink - Tracker: own design expected to be tested later this year - Spectrometer: we are building a second „hybrid“ with better options for temperature control

Current status of Pandora - Head sensor: very mature; gets a camera in the next version - Temperature controller: we are trying to improve the heat sink - Tracker: own design expected to be tested later this year - Spectrometer: we are building a second „hybrid“ with better options for temperature control - Software: Operation Software is on Version 1. 1; Processing software is not automized yet; both are written in Python

Current status of Pandora - Head sensor: very mature; gets a camera in the next version - Temperature controller: we are trying to improve the heat sink - Tracker: own design expected to be tested later this year - Spectrometer: we are building a second „hybrid“ with better options for temperature control - Software: Operation Software is on Version 1. 1; Processing software is not automized yet; both are written in Python - Have about 3 years of data with Pandora 1 (sun only), and about 8 months of data with two more units.

Current status of Pandora - Head sensor: very mature; gets a camera in the next version - Temperature controller: we are trying to improve the heat sink - Tracker: own design expected to be tested later this year - Spectrometer: we are building a second „hybrid“ with better options for temperature control - Software: Operation Software is on Version 1. 1; Processing software is not automized yet; both are written in Python - Have about 3 years of data with Pandora 1 (sun only), and about 8 months of data with two more units. - Only sun data are currently being analyzed (except for CINDI blind intercomparison)

Pandora direct sun data during CINDI Show plots of the same 11 days with good conditions for the parameters retrieved.

Pandora direct. Pandora sun data during CINDI NO 2 – calibrated OMI using MLE

MLE=Modified Langley Extrapolation Red: use mean over some measurements as a reference Blue: use convoluted extraterrestrial spectrum as reference

Pandora direct. Pandora sun data during CINDI O 3 – calibrated using OMI Brewer at GSFC

Pandora direct. Pandora sun data during CINDI H 2 O – calibrated. Aeronet using Aeronet at GSFC

Pandora direct sun data during CINDI The following species also „emerge“ from the algorithm. I don‘t know if any of these variations is real. . .

Pandora direct. Pandora sun data during CINDI HCHO – calibrated OMI using MLE

Pandora direct. Pandora sun data during CINDI Br. O – calibrated OMI using MLE

Pandora direct. Pandora sun data during CINDI Glyoxal – calibrated OMI using MLE

Pandora direct. Pandora sun data during CINDI SO 2 – calibrated OMI using MLE

Pandora direct. Pandora sun data during CINDI OCl. O – calibrated OMI using MLE

Pandora direct. Pandora sun data during CINDI IO – calibrated OMI using MLE