Differences in liquid cloud droplet effective radius and number concentration estimates between MODIS collections 5.1 and 6 over global oceans

Abstract

Abstract. Differences in cloud droplet effective radius and cloud droplet number concentration (CDNC) estimates inferred from the Aqua-MODIS (Moderate Resolution Imaging Spectroradiometer) collections 5.1 (C5.1) and 6 (C6) cloud products (MYD06) are examined for warm clouds over global oceans for the year 2008. Individual pixel level retrievals for both collections are aggregated to 1 • × 1 • and compared globally and regionally for the three main spectral channel pairs used for MODIS cloud optical property retrievals. Comparisons between both collections are performed for cases in which all three effective radii retrievals are classified by the MODIS cloud product as valid. The contribution to the observed differences of several key MYD06 Collection 6 algorithm updates are also explored, with a focus on changes to the surface reflectance model, assumed solar irradiance, above-cloud emission, cloud-top pressure (CTP), and pixel registration. Global results show a neutral to positive (> 50 cm −3 ) change for C6-derived CDNC relative to C5.1 for the 1.6 and 2.1 µm channel retrievals, corresponding to a neutral to −2 µm difference in droplet effective radius (r e ). For 3.7 µm retrievals, CDNC results show a negative change in the tropics, with differences transitioning toward positive values with increasing latitude spanning −25 to +50 cm −3 related to a +2.5 to −1 µm transition in effective radius. Cloud optical thickness (τ ) differences were small relative to effective radius and found to not significantly impact CDNC estimates. Regionally, the magnitude and behavior of the annual CDNC cycle are compared for each effective radius retrieval. Results from this study indicate significant inter-collection differences in aggregated values of effective radius due to changes to the precomputed retrieval lookup tables (LUTs) for ocean scenes, changes to retrieved cloud-top pressure, solar irradiance, or above-cloud thermal emission, depending upon spectral channel. The observed differences between collections may have implications for existing MODIS-derived climatologies and validation studies of effective radius and CDNC.

Remote Sensing of Droplet Number Concentration in Warm Clouds: A Review of the Current State of Knowledge and Perspectives

“…In agreement with the results found by Rausch et al (), N d tends to have lower values across the tropics and subtropics in C6 data compared to C5 (using the 3.7‐μm retrieved r e in both cases). In Rausch et al () N d differences were attributed to the corrections to the band‐averaged solar irradiances, atmospheric emission factors, and changes in cloud top pressure as used in the new C6 look up tables. On average, the root mean square differences among the three data sets examined are approximately within 20 cm −3 or ±20%.…”

“…Other improvements to the handling of the surface reflectance in Collection 6 (see Platnick et al, ) include a new surface spectral albedo data set derived from dynamic 8‐day sampling of MODIS data (previously Collection 5 used a 5‐year surface albedo climatology, see Platnick et al, ) and the inclusion of land spectral emissivities that are consistent with the cloud top property algorithm. Rausch et al () showed that for one MODIS ocean scene Collection 6 r e was up to 1 μm lower than that from Collection 5 for optically thin ( 2–3) clouds, which may be due to the surface albedo, although other changes between Collections 5 and 6 have also been made. However, Zhang and Platnick () found little effect upon r e3.7 versus r e2.1 differences from the implementation of the Cox and Munk () scheme for global marine liquid clouds.…”

Global and regional estimates of warm cloud droplet number concentration based on 13 years of AQUA-MODIS observations

“…These improvements include a better coregistration of the visible and near-infrared focal planes of the Aqua-MODIS instrument as well as significant improvements in the forward radiative transfer models used in the retrieval framework (Platnick et al, 2015). The impact of these changes on CDNC retrievals and gridded climatologies is assessed in detail in Rausch et al (2017). For the current study, Level-2 cloud retrievals are from Aqua-MODIS spanning the years 2003 through 2015.…”

Exploring Impacts of Size‐Dependent Evaporation and Entrainment in a Global Model

“…Figure 3a shows the effective radius of the cloud droplets in the top layer of stratiform clouds without ice in the default PD-simulation, while Figure 3b shows the difference in the effective radius between PI and PD. The effective radius of cloud droplets in CAM5.3-Oslo is slightly lower than that of MODIS (Moderate Resolution Imaging Spectroradiometer) (Rausch et al, 2017), which we find acceptable due to the positive biases associated with the r e -products coming from MODIS (Liang et al, 2015;Painemal & Zuidema, 2011). Changes in r e between PI and PD are mostly negative, owing to the Twomey effect when aerosol concentrations increase.…”