HIRDLS Ozone V003 (v2.04.09) Characteristics B. Nardi, C. Randall, V.L. Harvey & HIRDLS Team

HIRDLS Ozone V003 (v2.04.09) Characteristics

B. Nardi, C. Randall, V.L. Harvey & HIRDLS Team

HIRDLS Science MeetingBoulder, Jan 30, 2008

1 km

1 km

OZONESONDE Profile Comparisons

1 km

Fine vertical scale features are resolved.At highly variable early-spring high latitude, effect of sampling different air-masses is accentuated: smallest scales do not match.

High latitude

Mid-latitude

Low latitude

1 km

MLO[19.5oN, 156oW]

TMF[34.5oN, 118oW]

OZONE Lidar Profile Comparisons

1 km

OZONE Lidar comparisons – Statistical Differences

Mauna Loa Observatory [19.5oN, 156oW] Table Mountain Facility [34.5oN, 118oW]

HIRDLS bias is generally low, 5-10%, except region ~ 5-25 hPa, 5-10% high bias.

OZONESONDE Comparisons – Statistical Differences

INTEX-B AROTAL Lidar

Tight coincidence criteria (50km, 3hrs) shows low bias of <5%, except region 50-80 hPa, where it’s <10%.

MLS v2.2

Zonal Mean

2006Jul 2006Oct

2006Feb 2006May

O3 Percent Diff (%)O3 Percent Diff (%)

Pre

ssu

re (

hP

a)P

ress

ure

(h

Pa)

Pressure surfaces1 hPa 3 hPa

51 hPa 100 hPa

O3 Percent Diff (%) O3 Percent Diff (%)

31 hPa10 hPa

Lat

itu

de

Lat

itu

de

Lat

itu

de

Anticycloneborder

HIRDLSv2.04.09

MLSv2.2

Polar Vortexborder

Shown is the ozone standard deviation in different equivalent latitude and potential temperature bins. Results are in terms of percentage of mean VMR (black lines, 10%; white lines, 100%). This variability includes both the atmospheric variability and the random error of the HIRDLS data. The atmospheric variability is a minimum in summer, so the variability in high latitude summer is an upper bound of the HIRDLS precision. This is estimated at 2-8% between 500-2000K (1-50 hPa).

Ozone Estimated Precision (v2.04.09)

December 22June 20

~50 km, ~1 hPa --

~15 km, ~120hPa --

SUMMARY

1. An estimate of the HIRDLS ozone precision is 5-10% between 1-50 hPa, based on the variability of HIRDLS measurements in regions of minimum geophysical variability

2. HIRDLS ozone is reliable between 1 hPa – 100 hPa at mid and high latitudes and between 1 – 50 hPa at low latitudes.

3. HIRDLS ozone accuracy is better than 10% between 1 - 30 hPa (HIRDLS biased generally low). Some lidars and sondes indicate better accuracy: 5% between 2 - 10 hPa, and between 30-40 hPa, respectively.

4. A region of slightly positive 5% HIRDLS bias exists in a limited pressure range within 10-30 hPa at nearly all latitudes; this is observed by comparisons with sondes (SHADOZ, WAVES), lidars (MLO, TMF) and satellites (ACE, MLS).

5. At low latitudes a high HIRDLS bias begins at ~50 hPa and increases rapidly with increasing pressure. This may be caused by spikes indirectly related to the presence of clouds.

6. Ozonesonde and lidar profile comparisons give a strong indication that HIRDLS is capable of detecting fine vertical structure in the ozone field on the order of 1 to 2 km.

7. HIRDLS is capable of resolving low ozone pocket features associated with anticyclones in the highly variable northern winter high latitudes.

Odds & Ends

1. At mid and high latitude HIRDLS ozone may be reliable to several hundred hPa in the absence of local cloud features. Systematic bias is unknown due to lack of statistics;

2. The lower boundary for usable tropical lower stratospheric ozone may be extended earthward (~68 hPa) by using the HIRDLS 12.1 micron extinction parameter to screen data.

3. Negative values of retrieved total error ("O3Precision") indicate a >50% contribution to the error from a priori sources, and thus a strong a priori influence in the result. This is typically not an issue with tropical lower stratospheric ozone.

4. Upcoming release has increased vertical range, spaceward of 1 hPa, and at the earthward limit as well. The low bias is somewhat lessened, especially in the lower stratosphere.

5. A priority for future ozone improvements is in diminishing the effects of clouds in the ozone retrieval and screening remaining cloud at low latitudes.

HIRDLS Data Quality Document (Updated version to be posted very soon)5.2 OzoneSpecies Ozone (O3)Data Field Name: O3Useful Range: 1hPa – 100+ hPaVertical Resolution: 1.2 - 2 kmContact: Bruno NardiEmail: [email protected] Paper: Nardi et al., Initial Validation of Ozone Measurements from the High Resolution Dynamic Limb Sounder (HIRDLS), in review, J. Geophys. Res., 2007

HIRDLS Data Quality Document

5.2 Ozone

Species Ozone (O3)Data Field Name: O3Useful Range: 1hPa – 100+ hPaVertical Resolution: 1.2 - 2 kmContact: Bruno NardiEmail: [email protected] Paper: Nardi et al., Initial Validation of Ozone Measurements from the High Resolution Dynamic Limb Sounder (HIRDLS), in review, J. Geophys. Res., 2007

ACE-FTS (satellite solar occultation)

ACE satellite comparisons since May 2006 show agreement to within ~10% between 3 - 25 hPa in the NH (left) and between 1 - 50 hPa in the SH (right), with HIRDLS biased generally low.

Figure 5.2.4. A single orbit curtain plot comparison with MLS (Left: MLS; Middle: HIRDLS; Right: Difference) indicates HIRDLS is tends to be 10% lower earthward of ~30 hPa and is generally 10% high above ~1hPa and rapidly increases. HIRDLS ozone tends to be slightly lower than MLS, especially in the (night-time) descending node (RHS of panels).

Mean Difference1σ of DifferencesIndividual Differences.. ..

…..

Figure 5.2.1. Ozone difference between 97 SHADOZ Network (low latitude) ozonesonde profiles and 1042 coincident HIRDLS profiles, in terms of mixing ratio, left and in terms of percent (of sonde values), right. Shown are the mean difference (solid blue), the standard deviation (dashed blue) and the individual differences (black dots) from which these are derived.

SHADOZ Network

Figure 5.2.2. Ozone difference between 73 MLO lidar profiles and the 659 coincident HIRDLS profiles, in terms of mixing ratio, left, and percent (of sonde values), right.

MLO lidar

Figure 5.2.3. ACE satellite solar occultation comparisons since May 2006 show agreement to within ~10% (bottom) between 3 - 25 hPa in the Northern Hemisphere (left) and between 1 - 100 hPa in the Southern Hemisphere (right), with HIRDLS biased generally low.

ACE satellite

Figure 5.2.5. Mercator representations of the ozone percent difference between HIRDLS and collocated MLS (v2.2) for 2006-July-15, at pressure levels, 1 hPa, 3 hPa, 10 hPa, 31 hPa, 51 hPa and 100 hPa, as indicated over each sub-plot. The ozone fields for this day are relatively quiescent. Clearly illustrated in this and the previous figure is that in the tropics earthward of ~60 hPa HIRDLS ozone tends to be very high with respect to MLS. This is probably related to the presence of high cloud-tops.

MLS (v2.2)

Figure 5.2.6 Comparisons of HIRDLS profiles with: ozonesonde profile from La Reunion Island (left), and with MLO lidar profiles (right).

Anticycloneborder

HIRDLSv2.04.09

MLSv2.2

Polar Vortexborder

Ozone

Figure 5.2.7. An estimate of the HIRDLS ozone precision based on ozone variability in equivalent latitude bins. High precision estimates at winter-hemisphere high latitudes (LHS of top plots; RHS of bottom plots) may be indicative of the breakdown of the estimate rather than of an actual deterioration of HIRDLS precision.