Zur Seitenansicht
 

Titelaufnahme

Titel
Characterizing thin clouds using aerosol optical depth information
VerfasserStuke, Sophie
Betreuer / BetreuerinBlumthaler, Mario ; Kuhn, Michael
Erschienen2016
HochschulschriftInnsbruck, Univ., Masterarb., 2016
Anmerkung
Arbeit an der Bibliothek noch nicht eingelangt - Daten nicht geprüft
Datum der AbgabeNovember 2016
SpracheEnglisch
DokumenttypMasterarbeit
Schlagwörter (DE)Aerosole / Wolken / Wolken optische Dicke / Aerosol optische Dicke / Streuung
Schlagwörter (EN)clouds / aerosols / cloud optical depth / aerosol optical depth / scattering
URNurn:nbn:at:at-ubi:1-5092 Persistent Identifier (URN)
Zugriffsbeschränkung
 Das Werk ist frei verfügbar
Dateien
Characterizing thin clouds using aerosol optical depth information [1.49 mb]
Links
Nachweis
Klassifikation
Zusammenfassung (Deutsch)

-

Zusammenfassung (Englisch)

Based on direct solar irradiance measurements, this thesis tries to infer optical properties of clouds. Measurements have been collected in Innsbruck, Austria, using a precision filter radiometer (PFR) with four wavelength channels at 368, 412, 501 and 861 nm. Routinely, these measurements are used to infer aerosol properties, like aerosol optical depth (AOD) or the Angström exponent alpha. Operational aerosol retrievals are only possible during situations when no clouds are present in the field-of-view of the PFR. Therefore, only a limited number of direct solar irradiance measurements with clouds obscuring the sun (i.e., cloud-contaminated measurements) have been used for any analysis so far.

This unused dataset of cloud-contaminated measurements, taken from 2007 to 2014, is used to derive cloud optical depth (COD). In order to achieve this, a straightforward approach is employed: Extinction by NO2, O3 and Rayleigh scattering are taken into account in the retrieval of the optical depth (OD) for cloud-contaminated measurements. However, the derived OD is influenced by both the presence of aerosols and clouds obscuring the sun (i.e., OD = COD + AOD), while other contributors are neglected. The influence of the AOD on single measurements is estimated by using a daily mean value of AOD retrievals of the same day (gathered by the PFR during cloud-free conditions), or the previous day, if no AOD retrievals are available on the specific day. This approach is valid because variation of aerosol properties and cloud optical properties exhibit different time scales. The COD is calculated by subtracting the daily mean AOD value from single OD retrievals (i.e., COD = OD - mean(AOD)).

Derived COD values range from 0.10 to 2.64. Therefore, only optically thin clouds with a COD below three can be investigated using the proposed approach. The COD retrievals exhibit an uncertainty because of using the mean AOD in the derivation: the AOD (501 nm) can vary by up to 0.2 on specific days and employing the daily mean value of AOD does not account for this variability. Based on the spectral COD retrievals using the four channels of the PFR and employing Angström's turbidity formula, the Angström parameter alpha and the curvature parameter gamma are derived. The parameter alpha exhibits a value range from -0.82 to 1.99 with the 10th-percentile at -0.23 and the 90th-percentile at 0.10 for cloud-contaminated measurements. The curvature parameter gamma exhibits a value range from -3.02 to 6.52 with a mean of 0.02. However, the 10th-percentile is at -0.06 and the 90th-percentile at 0.09, i.e. for the majority of the retrievals gamma is nearly zero. Therefore, it is not further taken into account.

Based on hourly all-sky images, the cloud types obscuring the sun during specific measurements are determined for the period from 2011 to 2013 by manual inspection. However, events are only taken into account if (i) the solar disc is covered by a single cloud type, (ii) the solar zenith angle is not too high to avoid contamination by different cloud types along the slanted path through the atmosphere, and, (iii) at least two consecutive all-sky images show the same cloud type. Only Cirrus (Ci), Cirrostratus (Cs), Stratocumulus (Sc) and Altocumulus (Ac) clouds are analyzed in detail because for other cloud types too few events have been recognized (e.g., only 3 events for Stratus). Cirrus clouds show a COD range from 0.17 to 0.72 with an Angström parameter between -0.20 and -0.04. The other cloud types' value ranges for COD are comparable: 0.21 to 1.84 for Cs, 0.40 to 2.37 for Sc and 0.17 to 2.19 for Ac. Angström exponents range from -0.22 to 0.02 for Cs, -0.36 to 0.05 for Sc and -0.33 to -0.05 for Ac.

The analyzed ice clouds (Ci, Cs) and liquid water clouds (Sc, Ac) exhibit differences in the respective COD-alpha-relations. Values for Ci cluster in a narrow region for both parameters with a mean COD of 0.4 and a mean alpha of -0.1. Cirrostratus clouds show alpha values also with a mean of -0.1, but approximately constant over the whole COD range. However, water clouds (Sc, Ac) show a decreasing alpha with increasing COD, down to -0.3 for COD approximately equals 2.3. It is hypothesized, that the less negative Angström exponents of ice clouds arise from the weaker spectral dependence of the extinction efficiency for typical Ci and Cs ice crystal habits and sizes in the VIS and NIR spectral range. The spectral variability of extinction efficiency for water clouds is higher than for ice clouds in these wavelength ranges, yielding more negative alpha values. Moreover, multiple scattering within water clouds is hypothesized to further increase this difference.