The Observational Record Visual Cloud Observations

The longest record of cloudiness comes from synoptic reports made by human observers at weather stations over land and on ships over the ocean. These include the fraction of sky dome covered by all clouds and clouds in the lowest layer as well as morphological cloud types at low, mid, and high levels. The largest archive of synoptic cloud observations is the Extended Edited Cloud Report Archive (EECRA) (Hahn and Warren 1999). Several investigations have found that interannual variability and trends in surface-observed cloudiness are physically consistent with trends and variability in related meteorological parameters in some regions of the world (e.g., Norris 2000a, 2005b; Park and Leovy 2000; Sun et al. 2001). Although visual cloud observations lack quantitative radiative information, the application of locally derived linear empirical coefficients can nonetheless provide useful (but incomplete) information on changes in ERB. The reason for this is that monthly anomalies in cloud cover, the most quantitative parameter in synoptic cloud reports, are the largest contributors to monthly anomalies in ERB. The smaller radiative impacts of variations in cloud albedo, cloud emissivity, and cloud height (aside from differing changes in low- and upper-level cloud cover) cannot be derived from visual cloud observations.

Short- and longwave anomalies estimated from gridded surface cloud cover observations correspond substantially to satellite measurements of SWCRF and LWCRF on interannual and decadal timescales for many locations of the world (Norris 2005a). In regions with inadequate sampling, such as some tropical land areas and most tropical and southern hemisphere ocean areas, there is much less agreement. The limited information obtained from surface cloud observations is nonetheless valuable because it provides the only near-global record of cloud cover and related ERB changes during the pre-satellite era. Satellite measurements of upper-level cloud cover and LW radiation are consistent with the surface-observed record during the period of overlap, suggesting that it may also be reliable in the pre-satellite era (Norris 2005a). For example, the time series of upper-level cloud cover anomalies obtained from the EECRA and from the International Satellite Cloud Climatology Project (ISCCP) (Rossow and Schiffer 1999) coincide well (top of Figure 2.2), apart from the 1991-1993



Total -


gray = ISCCP

black = EECRA

I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1

1950 1960 1970 1980 1990 2000



Cu gray = ISCCP

black = EECRA

I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1

1950 1960 1970 1980 1990 2000

Figure 2.2 Time series of anomalies in total, upper-level, low-level, and cumulus cloud cover from ISCCP satellite (gray) and EECRA surface (black) observations averaged between 60°S and 60°N over land and ocean.

time period when aerosols resulting from the Mt. Pinatubo volcanic eruption caused upper-level cloudiness to be misidentifi ed as low-level cloudiness in ISCCP (Luo et al. 2002).

Surface-observed, upper-level cloud cover has declined globally over the past several decades (Figure 2.2), and this has weakened LWCRF over time. Although it appears that this long-term decrease in upper-level cloud cover was abruptly reversed in October 2001, the increase reported by ISCCP is actually an artifact. Use of an incorrect calibration coefficient for the NOAA-16 polar orbiter produced an underestimate of cloud-top temperature (Knapp 2008), which correspondingly caused an overestimate of upper-level cloud cover. Surface and satellite records disagree on trends in total cloud cover, low-level cloud cover (Figure 2.2), and reflected SW radiation. The surface data indicate that in recent decades there has been a large increase in low-level cloud amount, especially from cumulus clouds over tropical oceans. Such an enhancement of cloud cover would lead to a physically implausible reduction in absorbed SW radiation unless there was a large commensurate decrease in cloud albedo (a parameter that cannot be quantifi ed in visual cloud reports) (Norris 2005a). The origin of this apparent artifact has not yet been identified.

0 0

Post a comment