Assessment of the Usefulness of Atmospheric Satellite Sounder-Based Cloud Retrievals for Climate Studies Gyula I. Molnar Joel Susskind UMBC/JCET & NASA GSFC, Code 613, Greenbelt, MD 20771

Motivation: Global energy balance of the Earth- atmosphere system may change due to natural and man-made climate variations. For example, changes in the outgoing longwave radiation ( OLR) can be regarded as a crucial indicator of climate variations. Satellite SOUNDERS can retrieve OLR, as well as cloud-, and large-scale atmospheric variables [e.g., Temperature and moisture profiles] simultaneously.

 Satellite SOUNDER observations thus could help to improve the representation of climate feedbacks (profoundly affecting global climate change predictions) in climate models. The question also arises whether observed variabilities/trends in the longest-running satellite climatologies could be useful for a of ongoing climate change.  Satellite SOUNDER observations thus could help to improve the representation of climate feedbacks (profoundly affecting global climate change predictions) in climate models. The question also arises whether observed variabilities/trends in the longest-running satellite climatologies could be useful for a direct, observational detection of ongoing climate change.

TOVS TOVS Pathfinder Path-A The TOVS series of satellites were among the first to provide this type of information since 1979 OLR [Mehta and Susskind, 1999], cloud cover and cloud top pressure [Susskind et al., 1997] are among the key climatic parameters computed by the TOVS Pathfinder Path-A algorithm together with the retrieved temperature and moisture profiles. AIRS AIRS, regarded as the “new and improved TOVS”, can retrieve climatic parameters with a higher accuracy.

Are satellite climatologies “mature” enough for TREND assessments? 1)Here we present results that imply that the TOVS Path-A dataset is a good candidate 2)How do we substantiate this claim? 3)Cross-validations, e. g., OLR with CERES, clouds with ISCCP

It should be is an essential prerequisite for any dataset to accurately depict Interannual Variability in order to be considered a candidate for TREND assessments. Fortunately, even for rather short-term satellite datasets, cross-validation of mutually consistent depictions of interannual variability can be accomplished. The longest validated satellite dataset could then be used for trend assessment. The TOVS Path-A dataset is probably the longest, consistently produced dataset. So, let’s see what can it say about specific climatic variabilities/trends.

OVERVIEW OF TOVS Path-A RETRIEVAL METHODOLOGY Physically based system cloud-cleared Uses cloud-cleared radiances to produce solutions for atmospheric temperature and humidity profiles consistent Determines cloud parameters consistent with retrieved state and observed radiances; Computes (All-Sky) Outgoing Longwave Radiation [OLR], Clear- sky OLR from all parameters via radiative transfer.

Actual TOVS-retrieved parameters used in this work 1) P c A eff A c 1) The TOVS algorithm retrieves cloud top pressures [ P c ] and “effective” ( A eff, a product of infrared emissivity at 11  m and physical cloud cover or A c ) cloud fractions. 2) OLRClear-Sky OLR 2) OLR and Clear-Sky OLR 3) Temperature & humidity profiles 3) Temperature & humidity profiles NOTE: Most datasets are time-of-day [to 7:30 am] corrected

a) AIRS-MODIS cloud retrieval intercomparisons: We have shown earlier that there were significant differences between AIRS and MODIS cloud retrievals. However, when we created differences of monthly mean retrieval results to assess applicability of these cloud retrievals for climate variability studies, we see more reassuring results:

Note the general correspondence of the inter-annual variability patterns [despite that the AIRS mean cloud fraction is ~43% vs. ~67% of MODIS, whilst the AIRS mean cloud top pressure is ~520 mb vs. ~670 mb of MODIS], meaning that climatic trend-recognition could be close to identical for both retrieval schemes. This is further reinforced by the similar depiction of the much larger intra-seasonal differences, illustrated in the Fig. below showing August 2004 vs. January 2004 clouds fields and their differences [0.84 Corr. Coeff.].

B) ISCCP-TOVS Path-A Intercomparisons

Time series of Global cloud amount

Time series of Tropical cloud amount

Now, the actual Trends for this 16.0 yr {‘80-’00} Period: TOVS A eff : -1.0% ISCCP A c : -2.5% TOVS C tp : -4.1 mb ISCCP C tp : mb same Note: Almost the same trends for ANOMALIES

Seasonal Cycles of Global Mean Cloud Parameters

What is the latitudinal variability of these Trends? A) ISCCP D2 A c is all latitudes; B) TOVS A eff is slightly the Tropics; C) A c and A eff correlate the Tropics D) P c Trends anticorrelate in SH Tropics & 40 o N-70 o N, as well as 60 o S-80 o S.

Now, what about ISCCP-AVHRR TREND anomalies?

C) OLR Inter-comparisons: essentially an “energy-balance” The TOVS cloud-clearing is essentially an “energy-balance” approach, i. e., the [“observed” – “computed”] radiance errors are minimized with the retrieved/inferred cloud distribution. The “result” of this is clearly illustrated by the excellent correspondence between AIRS and CERES Outgoing Longwave Radiation values. The next 2 Figs. illustrate this between monthly mean OLR values as well as between their interannual variabilities.

Jan./2004-Jan./2003 CERES-AIRS Interannual OLR Variability

Global Mean Intercomparisons

OLR trends, exhibiting an ENSO-like pattern, generally anticorrelate with the corresponding TOVS A eff anomaly trend map. However, the TOVS C tp anomalies tend to correlate positively with the OLR anomalies.

Summary  AIRS and MODIS effective cloud fraction and cloud top pressure intra-seasonal and interannual anomalies correlate well enough to allow both datasets to be used for climate change assessments. This happens despite cloud top pressures of the AIRS (and of the ISCCP) clouds are consistently smaller than that of MODIS.  Likewise, ISCCP [in particular ISCCP-AVHRR] and TOVS Path-A cloud fraction and cloud top pressure interannual anomalies correlate well enough to allow these datasets to be considered for climate change studies.

 OLR interannual spatial variabilities from the available state-of-the-art CERES measurements and both from the AIRS and TOVS OLR computations are in remarkably good agreement, a very reassuring result regarding the overall quality of the AIRS and TOVS Path-A retrieval schemes.  Global mean TOVS OLR variability even in absolute terms is extremely close to that of the ‘state-of-the-art’ CERES values.

TOVS Path-A  We believe that studies like this could provide an impetus for a more ‘enthusiastic’ use of the existing satellite datasets for long- term climate analyses in general. We strongly encourage the use of the TOVS Path-A dataset due to its internal consistency (as well as due to the validation efforts described here).  We hope that the AIRS -based dataset will complement/continue these satellite- sounder-based climatologies well into the future.

Future Work A) The trends of rather short time series of climatic parameters can be significantly distorted by relatively short-period natural variability cycles [e. g., ENSO]. We plan to explore to find ways to minimize their effects on trends. For example, KNOWN, ENSO- affected areas/Months may be simply omitted from the trend-analysis database. B) Unify the TOVS & AIRS climate data time-series