Clarifying Remotely Retrieved Precipitation of Shallow Marine Clouds from the NSF/NCAR Gulfstream V
1. Introduction
Precipitation, a pronounced feature of the marine boundary layer, can hasten the stratocumulus-to-cumulus transition within both modeling and observational studies, by encouraging thermodynamically induced decoupling within the boundary layer (Paluch and Lenschow 1991; Sandu and Stevens 2011), aerosol depletion (Yamaguchi et al. 2017; O et al. 2018), and a more pronounced mesoscale organization (Blossey et al. 2021). Since precipitation is typically a poorly modeled quantity, robust observations of precipitation remain necessary for untangling subtle cause and effect relationships using modeling studies (e.g., vanZanten et al. 2011; Blossey et al. 2021). During the Cloud System Evolution in the Trades (CSET; Albrecht et al. 2019) campaign in 2015, precipitating stratocumulus and cumulus clouds were observed by in situ probes as well as by a 94-GHz Doppler radar, and a 532-nm-wavelength high spectral resolution lidar (HSRL) over the northeastern Pacific Ocean. Sarkar et al. (2020, hereafter SARKAR20), used these datasets to document the relationship of precipitation to the transition for three clear stratocumulus-to-cumulus transitions.
CSET was the first deployment upon the High-Performance Instrumented Airborne Platform for Environmental Research (HIAPER) Gulfstream V aircraft of both the HIAPER Cloud Radar (HCR) and the HSRL. The incorporation of the HCR into the campaign plan recognizes precipitation’s crucial role within the life cycle of shallow clouds. Cloud radars can sample large volumes of atmosphere and detect the large drops that are few enough to be missed by the in situ probes. Although rain, cloud liquid water and water vapor attenuate 94-GHz (3.2-mm-wavelength) radiation more than for 35-GHz (8.6-mm-wavelength) radars, the smaller 94-GHz antenna is physically more suitable for aircraft, and the shorter wavelength enhances their sensitivity to smaller drops. Aircraft cloud radars, in comparison to space-based radars (e.g., Kalmus and Lebsock 2016), also have the advantage that they can better resolve the near-surface precipitation giving rise to evaporation-driven density currents leading to further mesoscale cloud organization (Zuidema et al. 2017, and references therein). As such, 94-GHz radars have become a standard aircraft instrument for shallow cloud campaigns (Vali et al. 1998; vanZanten et al. 2005; Wood et al. 2011; Dzambo et al. 2019; Schwartz et al. 2019; Stevens et al. 2019; Pincus et al. 2021).
The CSET campaign documented the Lagrangian evolution of stratocumulus-to-cumulus clouds, including their precipitation, by sampling air parcels within the stratocumulus regime on the flight from California to Hawaii and resampling them approximately two days later on the way back to California. Near Hawaii, the clouds had evolved into cumuli with tops reaching approximately 2 km capable of supporting short-lived but intense rain burst events (Fig. 1). SARKAR20 document a shift toward larger drop sizes during the transition using two-dimensional cloud (2DC) probe measurements. Drops with diameters as large as 3.2 mm produced 1-s rain rates as high as 10 mm h−1 (240 mm day−1) during the cumulus rain bursts. This precipitation typically reached the near-surface (150-m) aircraft altitude, indicating the likelihood that cooling through evaporation and downdrafts of drier air from above could further support density-driven gust fronts.
(a),(b) Radar reflectivity, (c),(d) mean Doppler velocities and (e),(f) mean Doppler spectrum width as a function of longitude during the subcloud 150-m legs of (left) 17 Jul stratocumulus module RF06a and (right) 19 Jul cumulus module RF07c. The RF07c cumulus has evolved from the stratocumulus cloud sampled on 17 Jul as indicated by a HYSPLIT 500-m forward trajectory. The module labeling follows that within SARKAR20, in which the research flight number is appended by a cloud module letter in alphabetical order from east to west. The lidar-perceived cloud bases are shown as black dots and the flight position is shown with a blue line. The lowest available range gate (320 m for stratocumulus and 340 m for cumulus) is indicated with a red dashed line. Note the change in the y-axis range from left to right.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
Previous investigations of stratocumulus precipitation using 94-GHz radars have successfully applied simple power-law relationships between rain rate R and radar reflectivity Z (Z–R relationships) (vanZanten et al. 2005; Wood et al. 2011). The CSET campaign understood a priori that these might not apply for the deeper cumulus clouds because larger drop sizes would invalidate the retrievals. A combined radar–lidar retrieval of precipitation was developed for the moving platform by Schwartz et al. (2019), hereafter referred to as the (Z, β, v1) retrieval. The retrieval based on O’Connor et al. (2005), which was originally meant to retrieve rain parameters in stratocumulus regime, may or may not be correctly applicable to deeper precipitating stratocumulus and cumulus clouds as observed during CSET. The validity of using the (Z, β, v1) retrieval technique in a case study from CSET is discussed in this paper. The (Z, β, v1) method uses the ratio of the radar reflectivity (Z) to the lidar backscatter intensity (β), made proportional to the fourth power of the median volume diameter (Do) through a gamma-function drop size distribution assumption, to constrain the retrieved particle size.
The initial precipitation rate estimates based on the lidar/radar retrievals from the subcloud legs were an order of magnitude lower than the in situ values (Fig. 2h) with light precipitation rates not always retrieved (Fig. 2g). This is puzzling because the larger radar sampling volume should mean they are more likely to detect precipitation than the in situ probes (Wood 2005), not less. Subsequent improvements to the retrieval compare better to the in situ values, documented here in section 3, but the retrieved rain rates remain underestimated. This study seeks to identify further causes of the discrepancy. The examined hypotheses include the underlying drop size distribution assumption, rain and cloud liquid attenuation of the radar reflectivity, and Mie dampening of the radar reflectivity, contained in section 4.
(a) Z-weighted Mie-to-Rayleigh reflectivity ratio γ′, (c) effective diameter Deff, (e) total number concentration Nt, (g) rain rate R, vs longitude for the subcloud in situ (black) and from the (Z, β, v1) retrieval (blue), for stratocumulus cloud module RF06a on 17 Jul. (b),(d),(f),(h) As in (a), (c), (e), and (g), but for cumulus cloud module RF07c on 19 Jul. The remote retrievals are from an altitude of 340 m (380 m) for RF06a (RF07c) and the in situ values are from an altitude of 150 m. (a) shows only the in situ values as the (Z, β, v1) γ′ for stratocumulus were not available.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
An additional retrieval approach examines combining in situ information on the drop size distribution width with the radar reflectivity and mean Doppler velocity. The NSF/NCAR HIAPER plane, when deployed with the HCR, will almost certainly be accompanied by the in situ probes, because of plentiful space in the probe canisters and complementary applications. An example is the Organization of East Pacific Convection (OTREC) campaign held in 2019 (Fuchs-Stone et al. 2020). This is explored in section 5.
2. In situ assessment of the radar/lidar retrievals of precipitation microphysics
a. Observational datasets and retrieval methodology
The HCR is sensitive to −39.6 dBZ at a range of 1 km and can sample the boundary layer in both upward- and downward-pointing directions, although not simultaneously. The HSRL, operating at a 532-nm wavelength, can also point in either zenith or nadir directions, but not simultaneously. The two remote sensor datasets are placed on a common grid with a time resolution of 0.5 s, corresponding to a horizontal resolution of 50–100 m depending on aircraft speed, and a vertical resolution of 20 m (Schwartz et al. 2019). The most information can be derived from the remote sensors during the subcloud aircraft legs at approximately 150-m altitude, when the small distance to the low clouds reduces gaseous and liquid attenuation and the lidar can detect the cloud base independently of precipitation. The in situ data are gathered during the same subcloud legs and during the neighboring in-cloud-level legs. Both legs are typically 10 min in length, corresponding to a distance of approximately 70 km.
The 2DC optical array probes sample raindrops every second spanning diameters of 75–3200 µm, at 25-µm resolution (126 diameter bins total). Each individual image is quality controlled to obtain the drop size capable of producing a particular image [Korolev et al. (2007)], including a visual analysis of the preliminary in-field imagery for evidence of streaking. The rain rate is calculated using
Three independent measurements (radar reflectivity Z, lidar backscatter coefficient β, and the width of the radar Doppler velocity spectrum
The retrieval of Do primarily depends upon
b. Example comparison
The comparisons shown within Figs. 17 and 18 of SARKAR20, from a 10-min stratocumulus subcloud leg on 17 July (named module RF06a), and its air mass resampled as precipitating cumulus on 19 July (named module RF07c), are extended with a comparison of γ′ for RF07c in Fig. 2. The examined retrievals are from the same aircraft subcloud legs as the in situ measurements. Due to the radar-dead zone, the first-available radar range gate is at 340 m, approximately 200 m above the aircraft altitude of ~150 m.
Despite the increased potential for rain and cloud drop evaporation within the 200 m above the aircraft, and much smaller in situ sampling volumes, the 1-s in situ rain rates systematically exceed the remotely retrieved values, by approximately an order of magnitude. This discrepancy provided the original motivation for this study. In situ raindrop number concentrations (Nt) also exceed retrieved values by over an order of magnitude, while in situ Deff are slightly less than those retrieved. Although the Deff biases are smaller, they nevertheless far exceed the 14% error estimate from instrumental errors noted within O’Connor et al. (2005), in which the droplet size distribution shape is assumed known and Rayleigh scattering assumed for the radar (the Nt error is 33%, similarly calculated). Overall, the discrepancies cannot be explained by droplet growth through collision–coalescence within the intervening 200 m, because that process would diminish the in situ Nt and increase Deff. Even if only the difference in the rain rates is considered, collision–coalescence over 200 m cannot easily explain a discrepancy of an order of magnitude [see, e.g., discussion within Li et al. (2015)].
For the more heavily precipitating cumulus cloud, the retrieved rain frequencies slightly exceed those from the in situ probes, consistent with the smaller sampling volumes of the in situ probes. In contrast, precipitation is retrieved less frequently within the stratocumulus cloud than by the in situ probes. This is attributed to an algorithmic constraint for
c. Comparison of 10-min-averaged rain rates
Figure 3 compares the 10-min leg-mean rain rates [derived using the same protocol applied within Mohrmann et al. (2019)] to the (Z, β, v1) retrievals and the in situ values for the same 18 Lagrangian cases as examined within Mohrmann et al. (2019). Some air masses diverge into multiple locations, resulting in 30 cases in total. The retrieved 10-min leg-mean rain rate is constructed using the maximum rain rate within each vertical column, with the rain rate set to zero when no rain is detected. Using this averaging, Mohrmann et al. (2019) report CSET campaign-mean values of 2.2 and 0.7 mm day−1 for the California-to-Hawaii (outgoing) and Hawaii-to-California (incoming) flights (their Table 1). The calibration and water vapor attenuation corrections increase these values to 2.8 and 1.36 mm day−1. The lower in situ rain rates relative to the retrieved values, for rain rates < 0.01 mm day−1, can be explained by undersampling of larger raindrops by 2DC probes. The in situ rain rates continue to exceed the retrieved values for rain rates > 0.01 mm day−1 (Fig. 3). The average subcloud in situ precipitation rates for the CA-to-HI and HI-to-CA flights are 6.3 and 7.9 mm day−1, respectively. These increases by factors of 2.25 and 6 over the retrieved values also indicate a slight increase in the 10-min mean precipitation for the deeper clouds sampled closer to Hawaii. Although only reflecting a limited sample size, this also alters the perception of the hydrological cycle. When only those retrieved rain rates exceeding 0.01 mm day−1 are considered, the in situ rain rates averaged over both CA-to-HI and HI-to-CA flights increase to 21 mm day−1.
The 10-min-mean in situ rain rates within the near-surface 150-m-level legs for the same 18 Lagrangian cases examined within Mohrmann et al. (2019) vs the corresponding (Z, β, v1) retrieved rain rates computed following Mohrmann et al.’s (2019) algorithm, wherein the maximum retrieved rain rates within each vertical column are extracted for each 2-Hz sample and averaged for the 10 min. A one-on-one ratio line is included. The radar reflectivities have been corrected for a + 4.5 dB calibration offset and for water vapor attenuation, so that the remotely retrieved rain rates are slightly higher than those documented within Mohrmann et al. (2019). Some air masses of the nominal 18 Lagrangian cases diverge into two or more locations, resulting in 30 samples total.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
3. An improved (Z, β, v2) retrieval (version 2)
A subsequent version of the (Z, β) retrieval, hereafter referred to as (Z, β, v2), incorporated several improvements, listed below.
Improved calibration of the HCR based upon the ocean backscatter for each individual flight, along with an improved correction to the mean Doppler velocity, both done at the National Center for Atmospheric Research (NCAR). These corrections were already incorporated within SARKAR20 (a + 4.5 dB offset for RF07c) but are now extended to all of the flights. The calibration application includes a more thorough clutter identification applied to both the HSRL and HCR data, which also removes second trip echoes and sidelobes.
An accounting for water vapor attenuation using in situ and dropsonde observations following Fairall et al. (2018), based on the original formulation of Liebe et al. (1989), to all flights (also already incorporated within SARKAR20 for RF07c only).
A clearer articulation of the γ′ using smaller increments in the modal diameter and shape parameter, leading to a better match to Fig. 3 in O’Connor et al. (2005).
- A normalized drizzle number concentration Nw calculated from the measured radar reflectivity after accounting for Mie-to-Rayleigh reflectivity ratio γ′ [missing in Eq. (5) within Schwartz et al. (2019)] according to
- to which the total number concentration Nt corresponds to as
The contribution of the aircraft motion to the width of the Doppler spectra (contained within the turbulence term σt) is now the moving average based on approximately 20 samples of the aircraft speed (a distance of 1–2 km, better matching the boundary layer depth), as opposed to one only [Eq. (1) of Schwartz et al. 2019]. This recognizes the approximate scaling of the turbulence contribution with the boundary layer depth.
A mistake in the range gate start value was rectified, resulting in a reduced radar “dead zone” of 300 m, from a previous 400 m. The first usable height remains the dead zone + the GV altitude.
The improved retrievals (Z, β, v2) produce rain rates that are higher than those from the original (Z, β, v1) values, though still less than the in situ values (Fig. 4a). The (Z, β, v2) median rain rate is 16 mm day−1 for the same subcloud RF07c leg shown in Fig. 2. The comparable median rain rate is 2 mm day−1 for the (Z, β, v1) retrieval, and 44 mm day−1 based on the in situ data. The raindrop number concentration also matches the in situ measurements more closely than the original retrieval (Fig. 4b), if still biased low. However, the retrieved Deff has decreased significantly, worsening the comparison (Fig. 4c). Consistent with the Deff underestimate, γ′ remains overestimated (Fig. 4d). The retrieved shape parameter is too large (Fig. 4e), overly narrowing the raindrop size distribution compared to the in situ size distribution.
(a) Rain rates, (b) raindrop concentration, (c) effective diameter, (d) Mie-to-Rayleigh reflectivity ratio γ′, and (e) μg for (Z, β, v1) and (Z, β, v2) at ~300-m altitude, and in situ observations at 150-m altitude as a function of longitude, for cloud module RF07c. The aircraft-measured in situ vertical velocities are also shown as a function of longitude at (f) 150 m and the nearest (g) 600-m altitude flight legs, where the positive values denote updrafts.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
4. Potential causes for the retrieval bias
The improvements incorporated into (Z, β, v2) produce improved rain rates, but systematic biases remain. Here we discuss three potential reasons: 1) the underlying assumption of a gamma distribution, 2) attenuation by rain and liquid, and 3) difficulty in fully incorporating γ′ within the iterative retrieval.
a. Assessment of the gamma functional fit
Gamma distribution fits are applied to the in-cloud and subcloud drop size distributions of both stratocumulus and cumulus clouds of four to six cloud modules (Fig. 5, also examined in Fig. 12 of SARKAR20). Inverse exponential fits truncated at 76 µm and 3.2 mm are also shown. The gamma functions adequately represent each of the stratocumulus drop size distribution, and the in-cloud cumulus drop size distribution. The median volume diameter Do increases from 0.5 mm within stratocumulus to 0.8 mm below the cloud. Do within cumulus is larger at 1.1 mm and increases to 1.3 mm below the cloud. The shape parameter μg is more negative for stratocumulus (−1.9 at 150 m and −2.5 at in-cloud level) than for the cumulus clouds (0 at 150 m and −1.2 at in-cloud level), indicating a shift to a narrower size distribution centered at bigger drop sizes for cumulus. This is consistent with Geoffroy et al. (2014), who derived μg (their υ − 1) values of 0–2 from in situ microphysical measurements from the Rain in Cumulus over Ocean campaign (Rauber et al. 2007; Zuidema et al. 2012b). The gamma fit to the subcloud cumulus cloud drop size distribution, the condition with the most intense shallow rain, is weaker than for the other three cases. Here, the truncated exponential fit performs best. This suggests the truncated exponential might produce larger rain rates and Deff values overall, if applied within the retrieval, and supports its application for the global CloudSat precipitation retrievals (Lebsock and L’Ecuyer 2011). The lognormal fit [also applied within vanZanten et al. (2005)] is also adequate and arguably provides the best compromise for depicting both light and more intense shallow precipitation.
The 1-Hz in situ 2DC raindrop size distributions (dN/dD) for stratocumulus clouds measured (a) below cloud, at ~150 m, and (b) within cloud. (c),(d) As in (a) and (b), but for cumulus clouds. Lognormal (red), truncated exponential (at 76 µm and 3.2 mm; dark blue), and gamma (light blue) distribution fits are indicated in each panel along with the values of the gamma fit parameters. In situ data drawn from the same cloud modules assessed in Fig. 12 of SARKAR20.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
b. Assessment of attenuation from cloud and rain
Another hypothesis for the retrieval bias within cumulus clouds is a lack of correction for radar attenuation by liquid and rainwater, potentially reducing radar sensitivity to the scattering by the target of interest (Lhermitte 1989; Pujol et al. 2006; Matrosov 2007; Chandra et al. 2015; Fairall et al. 2018; Oh et al. 2020). The attenuation can be notable for millimeter-wavelength radars (Matrosov 2007), and downward-looking W-band radars often show a reduction in reflectivity near cloud bases from either attenuation, or raindrop evaporation, creating an ambiguity (Mason et al. 2017). The path-integrated attenuation is large enough to underpin its own rain-rate retrieval (L’Ecuyer and Stephens 2002; Fairall et al. 2018).
Effects of attenuation by both cloud liquid and rain are analyzed. Although a precise approach for estimating cloud liquid attenuation would require a dual-wavelength radar (e.g., Hogan et al. 2005), an attenuation estimate can be made for the cumulus case shown here using the mean in-cloud liquid water content and mean cloud thickness. The two-way liquid water attenuation coefficient in dB km−1 is proportional to the cloud liquid water content (LWC) (Atlas and Ulbrich 1977; Pujol et al. 2006):
The rain attenuation is larger, but nevertheless remains on par with the uncertainty of 1–2 dB associated with the radar reflectivity (Romatschke et al. 2021). This is demonstrated through an estimation based on the vertical profiles of radar reflectivity. For the subcloud legs with an upward-viewing radar,any reduction in reflectivity between the radar and the altitude of maximum reflectivity (typically below 400-m altitude) is assumed to be from raindrop evaporation only. Reduction in the reflectivity above this altitude is assumed to be from signal attenuation. An attenuation correction is only applied when the column maximum reflectivity exceeds 10 dBZ and the Doppler velocity is higher than 1 m s−1, at or below 1-km altitude. These thresholds are determined from the 1-Hz in situ rain rates greater than 1 mm h−1, but are similar to those applied within Chandra et al. (2015) and Oh et al. (2020); 70% (257 out of 371) of the radar profiles for RF07c satisfied this criteria. The two-way attenuation is thereafter calculated following Fairall et al. (2018):
For the RF07c cumulus example shown here, the attenuation correction increases the reflectivity by 0.5–1 dBZ in the lower 1 km of the cloud, and by 1–3 dBZ near the cloud tops (not shown). The RF07c cumulus module was selected in part because of its intense rain shower. These attenuation values represent upper limits on the potential impact from rain attenuation upon the measured radar reflectivity. Given the radar reflectivity is uncertain by 1–2 dB, attenuation by rain can be ruled out as a significant contributor to the retrieval bias, particularly below the cloud.
c. Underacknowledgement of Mie dampening within the (Z, β) retrieval
The need for an additional constraint on particle size is a key motivation for incorporating the independent lidar backscatter measurement into the (Z, β) retrievals; the Mie-induced dampening (Mie 1908) of the reflectivity Z for the larger raindrops may dominate the rain rates. Figures 6 and 7 reinforce the inordinate impact of the frequently occurring large raindrops (>200 µm) upon the radar reflectivity. A neglect of Mie scattering in an R = aZb relationship can increase the sensitivity of reflectivity-derived rain-rate estimates to any errors in the radar reflectivity (Fig. 6), and underestimate rain rates by 50%–90% for drops larger than 200 µm [Fig. 7; calculated using the Mätzler (2002) algorithm]. This contrasts with some previous work on stratocumulus clouds, for which traditional R = aZb relationships, with a and b empirically determined using microphysical datasets, work well (vanZanten et al. 2005; Wood et al. 2011). For those campaigns, the stratocumulus rain rates were much lighter, at approximately 0.5 mm day−1, than those measured during CSET.
(a) The 1-Hz reflectivity (dBZ) vs rain rate (mm day−1) with (red) and without (blue) Mie scattering accounted for, from the subcloud and in-cloud legs of Hawaii-to-California flights on 19 and 29 Jul and 9 Aug. (b) As in (a), but averaged over 10-min (~80-km spatial scale), for both subcloud and in-cloud legs from the 17–19 Jul, 27–29 Jul, and 7–9 Aug flight pairs. Only 1-Hz rain rates > 0.01 mm h−1 (0.24 mm day−1) are considered.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
Rain rates estimated from a Z–R relationship (Restimated) vs in situ rain rate (Rin-situ) for the Hawaii-to-California flights on 19 Jul, 29 Jul and 9 Aug. Empirical parameters a and b within Restimated = aZb are derived for each flight, for the subcloud (blue) and in-cloud (red) legs. The Z–R relationship is then applied to an in situ calculated radar reflectivity that also accounts for Mie scattering. The Mie scattering is accounted for using Matzler’s (2002) algorithm.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
The O’Connor et al. (2005) retrieval of rain rate was originally intended for the stratocumulus regime, for which Mie dampening of the reflectivity is typically not significant. Perhaps because of the original application, the γ′ parameter is determined after Do, μg, and Nw have been established iteratively. Figure 4a indicates the upper limit to the retrieved rain rates coincides with too-small Deff, producing a γ′ that is barely modulated from its initially assumed value of 1. We argue that additional lack of information on this constraint, which is difficult to estimate from the remote measurements, contributes to the underestimated rain rates shown in Fig. 4a.
5. A retrieval approach based on Z and
The suite of instrumentation typically deployed upon the NSF/NCAR Gulfstream V for precipitation studies will include both the HCR and the in situ microphysical probes, because the instruments complement each other and the space is readily available upon the GV for the probes. This suggests another retrieval approach for constraining the particle size, by combining the in situ droplet size distribution information with the radar reflectivity Z and the mean Doppler velocity
A full exploration of the relative strengths of a rain-rate retrieval based on
The Z-weighted mean terminal fall velocity (υd,lookup) is obtained from the in situ data in terms of the lognormal distribution
In Figs. 8 and 9, σ is held constant at 0.6, and a small (<1 dBZ) attenuation correction is also applied to the reflectivity. The rain rate, Deff, raindrop concentration and γ′ are computed using the retrieved μl and Nt through integrating over the moments of the lognormal distribution N(D). The
Longitude–height cross section for (Z, υd) and (Z, β, v2) are shown for (a),(b) rain rate, (d),(e) raindrop number concentration, (g),(h) Deff, and (j),(k) γ′ for the 19 Jul RF07c cumulus example. The aircraft position is indicated as a black dashed line. The 10-min leg-mean retrievals for (c) rain rate, (f) raindrop number concentration, (i) effective diameter, and (l) γ′, and in situ values from the nearest vertical leg [location shown in inset within (b)]. Dotted lines are cloud-only (>−40 dBZ) averages and full lines include both cloudy and clear samples.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
(a) Rain rates, (b) raindrop concentration, (c) effective diameter, and (d) Mie-to-Rayleigh reflectivity ratio γ′ at ~300-m altitude for (Z, β, v2), (Z, υd) retrievals and at ~150-m altitude for in situ observations as a function of longitude for cloud module RF07c.
Citation: Journal of Atmospheric and Oceanic Technology 38, 9; 10.1175/JTECH-D-20-0166.1
The
Comparison of the rain rates, rain number concentrations, effective diameter, and the Z-weighted γ′ parameter for the in situ, (Z, β) v2 and
A more comprehensive application to the full CSET radar dataset would benefit from incorporation of a vertically varying value for σ that evolves with boundary layer depth, to better capture size sorting effects on the distribution width. The propagation of errors upon the retrieval uncertainties through an optimal estimation framework also remains to be done. A more complex retrieval could also incorporate the Doppler spectrum width.
6. Summary and discussion
In situ rain rates often exceeded 1 mm h−1 (24 mm day−1) in cumulus clouds of 2-km height during CSET (SARKAR20), significantly higher than the initial rain rates retrieved using a combination of the HCR and HSRL (Schwartz et al. 2019; Bretherton et al. 2019; Mohrmann et al. 2019). The combination of radar and lidar measurements for rain-rate retrievals was originally proposed by O’Connor et al. (2005) for light stratocumulus drizzle regime, with rain rates below 1 mm h−1. However, this method is incorporated by Schwartz et al. (2019) for the deeper stratocumulus and cumulus precipitation observed during CSET, and it poses certain underestimates in the rain-rate retrievals as shown in the present study. Subsequent improvements to the Schwartz et al. (2019) retrieval increase the retrieved rain rates, if still less than the in situ rain rates. Three potential sources for the retrieval bias are explored: the gamma function fit for the drop size distribution, attenuation by rain and cloud water, and treatment of Mie dampening of the radar reflectivity within the retrieval. An exponential fit truncated for drops larger than 75 µm and a lognormal fit provides an improved fit to the drop size distribution beneath the cumulus cloud example of an intense, short-lived shower, although its implementation into the retrieval is not explored here. Attenuation by rain and cloud liquid is either negligible or comparable to the measurement uncertainty of the radar, keeping in mind that the comparisons are only applied to the subcloud legs. Last, the retrieval does not account for the dampening of the radar reflectivity by Mie scattering well, with the modeled raindrop size distribution typically too narrowly peaked (as indicated by a too-large μg) around too-small drop sizes. Also, γ′ becomes an additional free parameter that is difficult to constrain well with the remote measurements.
An alternative approach is explored based on the mean radar Doppler velocity
The NSF/NCAR Gulfstream V is a permanent platform for the requestable HCR and HSRL. The NSF/NCAR Gulfstream V research plane carried the HCR and HSRL and in situ probes for the Southern Ocean Clouds, Radiation, Aerosol Transport Experimental Study (McFarquhar et al. 2021). The plane’s deployment for the Organization of east Pacific Convection campaign included the HCR and the 2DC cloud probes, but not the HSRL (Fuchs-Stone et al. 2020). A small microwave radiometer was initially incorporated into the CSET instrument payload, to constrain the column liquid water (Zuidema et al. 2012a). Although the radiometer did not function on this deployment, it can provide an additional independent piece of information helpful to the retrievals. Further effort can also be placed into improving the lidar discrimination of the rain from the aerosol below cloud, using a priori information to better estimate the relative contributions to the backscattered intensity. The NCAR Gulfstream V plane is one of NSF’s premier platforms for atmospheric research, and the commitment to a repeated use of the same instrumentation indicates the value of deepening the characterization of their dedicated retrievals.
Acknowledgments
MS and PZ acknowledge support from NSF Grant AGS-1445832. VG was supported by the National Science Foundation (NSF) Grant AGS-1445831 awarded to the University of Chicago, the National Aeronautics and Space Administration (NASA) Grant 18-ACMAP18-0110 awarded to the Argonne National Laboratory, and the U.S. Department of Energy’s (DOE) Atmospheric System Research (ASR), an Office of Science, Office of Biological and Environmental Research (BER) program, under Contract DE-AC02-06CH11357 awarded to Argonne National Laboratory. We gratefully acknowledge the computing resources provided on Bebop, a high-performance computing cluster operated by the Laboratory Computing Resource Center (LCRC) at the Argonne National Laboratory. We thank two anonymous reviewers for their thoughtful comments, which made this a more valuable study.
Data availability statement
The version 2 merged, quality-controlled radar–lidar 2Hz dataset placed on a uniform georeferenced grid and (Z, β) v2 retrievals are available at https://doi.org/10.26023/WB9J-MFRK-QY0Z. More details on the retrievals are also available at https://data.eol.ucar.edu/datafile/nph-get/487.047/drizzle_ret_data_writeup.pdf. Details on the calibration offsets are available at https://doi.org/10.5065/D6CJ8BV7. All other datasets are available through the NCAR EOL archive at https://data.eol.ucar.edu/master_lists/generated/cset/.
REFERENCES
Albrecht, B., and Coauthors, 2019: Cloud System Evolution in the Trades (CSET): Following the evolution of boundary layer cloud systems with the NSF-NCAR GV. Bull. Amer. Meteor. Soc., 100, 93–121, https://doi.org/10.1175/BAMS-D-17-0180.1.
Atlas, D., and C. W. Ulbrich, 1977: Path- and area-integrated rainfall measurement by microwave attenuation in the 1–3 cm band. J. Appl. Meteor., 16, 1322–1331, https://doi.org/10.1175/1520-0450(1977)016<1322:PAAIRM>2.0.CO;2.
Beard, K. V., 1985: Simple altitude adjustments to raindrop velocities for Doppler radar analysis. J. Atmos. Oceanic Technol., 2, 468–471, https://doi.org/10.1175/1520-0426(1985)002<0468:SAATRV>2.0.CO;2.
Blossey, P. N., C. S. Bretherton, and J. Mohrmann, 2021: Simulating observed cloud transitions in the northeast Pacific during CSET. Mon. Wea. Rev., 149, 2633–2658, https://doi.org/10.1175/MWR-D-20-0328.1.
Bretherton, C. S., and Coauthors, 2019: Cloud, aerosol, and boundary layer structure across the northeast Pacific stratocumulus–cumulus transition as observed during CSET. Mon. Wea. Rev., 147, 2083–2103, https://doi.org/10.1175/MWR-D-18-0281.1.
Chandra, A., C. Zhang, P. Kollias, S. Matrosov, and W. Szyrmer, 2015: Automated rain rate estimates using the Ka-band ARM zenith radar (KAZR). Atmos. Meas. Tech., 8, 3685–3699, https://doi.org/10.5194/amt-8-3685-2015.
Duncan, D. I., P. Eriksson, S. Pfreundschuh, C. Klepp, and D. C. Jones, 2019: On the distinctiveness of observed oceanic raindrop distributions. Atmos. Chem. Phys., 19, 6969–6984, https://doi.org/10.5194/acp-19-6969-2019.
Dzambo, A. M., T. L’Ecuyer, O. O. Sy, and S. Tanelli, 2019: The observed structure and precipitation characteristics of southeast Atlantic stratocumulus from airborne radar during ORACLES 2016–17. J. Appl. Meteor. Climatol., 58, 2197–2215, https://doi.org/10.1175/JAMC-D-19-0032.1.
Evans, K. F., S. J. Walter, A. J. Heymfield, and G. M. McFarquhar, 2002: Submillimeter-wave cloud ice radiometer: Simulations of retrieval algorithm performance. J. Geophys. Res., 107, 4028, https://doi.org/10.1029/2001JD000709.
Fairall, C. W., S. Y. Matrosov, C. R. Williams, and E. J. Walsh, 2018: Estimation of rain rate from airborne Doppler W-band radar in CalWater-2. J. Atmos. Oceanic Technol., 35, 593–608, https://doi.org/10.1175/JTECH-D-17-0025.1.
Fang, M., B. Albrecht, E. Jung, P. Kollias, H. Jonsson, and I. PopStefanija, 2017: Retrieval of vertical air motion in precipitating clouds using Mie scattering and comparison with in situ measurements. J. Appl. Meteor. Climatol., 56, 537–553, https://doi.org/10.1175/JAMC-D-16-0158.1.
Frisch, A. S., C. W. Fairall, and J. B. Snider, 1995: Measurement of stratus cloud and drizzle parameters in ASTEX with a Ka-band Doppler radar and a microwave radiometer. J. Atmos. Sci., 52, 2788–2799, https://doi.org/10.1175/1520-0469(1995)052<2788:MOSCAD>2.0.CO;2.
Fuchs-Stone, Z., D. J. Raymond, and S. Sentic, 2020: OTREC2019: Convection over the east Pacific and southwest Caribbean. Geophys. Res. Lett., 47, e2020GL087564, https://doi.org/10.1029/2020GL087564.
Geoffroy, O., A. P. Siebesma, and F. Burnet, 2014: Characteristics of the raindrop distributions in RICO shallow cumulus. Atmos. Chem. Phys., 14, 10 897–10 909, https://doi.org/10.5194/acp-14-10897-2014.
Hogan, R., N. Gaussiat, and A. Illingworth, 2005: Stratocumulus liquid water content from dual-wavelength radar. J. Atmos. Oceanic Technol., 22, 1207–1218, https://doi.org/10.1175/JTECH1768.1.
Kalmus, P., and M. Lebsock, 2016: Correcting biased evaporation in CloudSat warm rain. IEEE Trans. Geosci. Remote Sens., 55, 6207–6217, https://doi.org/10.1109/TGRS.2017.2722469.
Khain, A. P., and Coauthors, 2015: Representation of microphysical processes in cloud-resolving models: Spectral (bin)microphysics versus bulk parameterization. Rev. Geophys., 53, 247–322, https://doi.org/10.1002/2014RG000468.
Korolev, A. V., G. A. Isaac, J. W. Strapp, S. G. Cober, and H. W. Barker, 2007: In situ measurements of liquid water content profiles in midlatitude stratiform clouds. Quart. J. Roy. Meteor. Soc., 133, 1693–1699, https://doi.org/10.1002/qj.147.
Lebsock, M. D., and T. S. L’Ecuyer, 2011: The retrieval of warm rain from CloudSat. J. Geophys. Res., 116, D20209, https://doi.org/10.1029/2011JD016076.
L’Ecuyer, T. S., and G. L. Stephens, 2002: An estimation-based rainfall retrieval algorithm for attenuating spaceborne radars. J. Appl. Meteor. Climatol., 41, 272–285, https://doi.org/10.1175/1520-0450(2002)041<0272:AEBPRA>2.0.CO;2.
Lhermitte, R., 1989: Attenuation and scattering of millimeter wavelength radiation by clouds and precipitation. J. Atmos. Oceanic Technol., 7, 464–479, https://doi.org/10.1175/1520-0426(1990)007<0464:AASOMW>2.0.CO;2.
Li, Z., P. Zuidema, P. Zhu, and H. Morrison, 2015: The sensitivity of simulated shallow cumulus convection and cold pools to microphysics. J. Atmos. Sci., 72, 3340–3355, https://doi.org/10.1175/JAS-D-14-0099.1.
Liebe, H. J., T. Manabe, and G. A. Hufford, 1989: Millimeter-wave attenuation and delay rates due to fog/cloud conditions. IEEE Trans. Antennas Propag., 37, 1617–1612, https://doi.org/10.1109/8.45106.
Mason, S. L., J. C. Chiu, R. J. Hogan, and L. Tian, 2017: Improved rain rate and drop size retrievals from airborne Doppler radar. Atmos. Chem. Phys., 17, 11 567–11 589, https://doi.org/10.5194/acp-17-11567-2017.
Matrosov, S. Y., 2007: Potential for attenuation-based estimations of rainfall rate from CloudSat. Geophys. Res. Lett., 34, L05817, https://doi.org/10.1029/2006GL029161.
Mätzler, C., 2002: MATLAB functions for Mie scattering and absorption. Universitas Bernensis Institut für Angewandte Physik Research Rep. 2002-11, 24 pp.
McFarlane, S. A., K. F. Evans, and A. S. Ackerman, 2002: A Bayesian algorithm for the retrieval of liquid water cloud properties from microwave radiometer and millimeter radar data. J. Geophys. Res., 107, 4317, https://doi.org/10.1029/2001JD001011.
McFarquhar, G. M., and Coauthors, 2021: Observations of clouds, aerosols, precipitation, and surface radiation over the Southern Ocean: An overview of CAPRICORN, MARCUS, MICRE, and SOCRATES. Bull. Amer. Meteor. Soc., 102, E894–E928, https://doi.org/10.1175/BAMS-D-20-0132.1.
Mie, G., 1908: Beitrage optik trüber medien, speziell kolloidaler metallösungen (Contributions to the optics of diffuse media, especially colloid metal solutions). Ann. Phys., 330, 377–445, https://doi.org/10.1002/andp.19083300302.
Mohrmann, J., and Coauthors, 2019: Lagrangian evolution of the northeast Pacific marine boundary layer structure and cloud during CSET. Mon. Wea. Rev., 147, 4681–4700, https://doi.org/10.1175/MWR-D-19-0053.1.
O, K.-T., R. Wood, and C. S. Bretherton, 2018: Ultraclean layers and optically thin clouds in the stratocumulus-to-cumulus transition. Part II: Depletion of cloud droplets and cloud condensation nuclei through collision–coalescence. J. Atmos. Sci., 75, 1653–1673, https://doi.org/10.1175/JAS-D-17-0218.1.
O’Connor, E. J., R. J. Hogan, and A. J. Illingworth, 2005: Retrieving stratocumulus drizzle parameters using Doppler radar and lidar. J. Appl. Meteor., 44, 14–27, https://doi.org/10.1175/JAM-2181.1.
Oh, S.-B., P. Kolias, J.-S. Lee, S.-W. Lee, Y. H. Lee, and J.-H. Jeong, 2020: Rain-rate estimation algorithm using signal attenuation of Ka-band cloud radar. Meteor. Appl., 27, e1825, https://doi.org/10.1002/met.1825.
Paluch, I. R., and D. H. Lenschow, 1991: Stratiform cloud formation in the marine boundary layer. J. Atmos. Sci., 48, 2141–2158, https://doi.org/10.1175/1520-0469(1991)048<2141:SCFITM>2.0.CO;2.
Pincus, R., and Coauthors, 2021: Observations from the NOAA P-3 aircraft during ATOMIC. Earth Syst. Sci. Data, 13, 3281–3296, https://doi.org/10.5194/essd-13-3281-2021.
Pujol, O., J.-F. Georgis, L. Feral, and H. Sauvageot, 2006: Degradation of radar reflectivity by cloud attenuation at microwave frequency. J. Atmos. Oceanic Technol., 24, 640–657, https://doi.org/10.1175/JTECH1992.1.
Rauber, R. M., and Coauthors, 2007: Rain in Shallow Cumulus over the Ocean: The RICO Campaign. Bull. Amer. Meteor. Soc., 88, 1912–1928, https://doi.org/10.1175/BAMS-88-12-1912.
Romatschke, U., M. Dixon, P. Tsai, E. Loew, J. Vivekanandan, J. Emmett, and R. Rilling, 2021: The NCAR airborne 94-GHz cloud radar: Calibration and data processing. Data, 6, 66, https://doi.org/10.3390/data6060066.
Sandu, I., and B. Stevens, 2011: On the factors modulating the stratocumulus to cumulus transitions. J. Atmos. Sci., 68, 1865–1881, https://doi.org/10.1175/2011JAS3614.1.
Sarkar, M., P. Zuidema, B. Albrecht, V. Ghate, J. Jensen, J. Mohrmann, and R. Wood, 2020: Observations pertaining to precipitation within the northeast Pacific stratocumulus-to-cumulus transition. Mon. Wea. Rev., 148, 1251–1273, https://doi.org/10.1175/MWR-D-19-0235.1.
Schwartz, M. C., and Coauthors, 2019: Merged cloud and precipitation dataset from the HIAPER-GV for the Cloud System Evolution in the Trades (CSET) campaign. J. Atmos. Oceanic Technol., 36, 921–940, https://doi.org/10.1175/JTECH-D-18-0111.1.
Sloss, P. W., and D. Atlas, 1968: Wind shear and reflectivity gradient effects on Doppler radar spectra. J. Atmos. Sci., 25, 1080–1089, https://doi.org/10.1175/1520-0469(1968)025<1080:WSARGE>2.0.CO;2.
Stamnes, K., S.-C. Tsay, W. Wiscombe, and K. Jayaweera, 1988: Numerically stable algorithm for discrete-ordinate-method radiative transfer in multiple scattering and emitting layered media. Appl. Opt., 27, 2502–2509, https://doi.org/10.1364/AO.27.002502.
Stevens, B., and Coauthors, 2019: A high-altitude long-range aircraft configured as a cloud observatory: The NARVAL expeditions. Bull. Amer. Meteor. Soc., 100, 1061–1077, https://doi.org/10.1175/BAMS-D-18-0198.1.
Vali, G., R. D. Kelly, J. French, S. Haimov, and D. Leon, 1998: Finescale structure and microphysics of coastal stratus. J. Atmos. Sci., 55, 3540–3564, https://doi.org/10.1175/1520-0469(1998)055<3540:FSAMOC>2.0.CO;2.
vanZanten, M. C., B. Stevens, G. Vali, and D. H. Lenschow, 2005: Observations of drizzle in nocturnal marine stratocumulus. J. Atmos. Sci., 62, 88–106, https://doi.org/10.1175/JAS-3355.1.
vanZanten, M. C. B., and Coauthors, 2011: Controls on precipitation and cloudiness in simulations of trade-wind cumulus as observed during RICO. J. Adv. Model. Earth Syst., 3, M06001, https://doi.org/10.1029/2011MS000056.
Williams, C. R., 2002: Simultaneous ambient air motion and raindrop size distributions retrieved from UHF vertical incident profiler observations. Radio Sci., 37, 1–22, https://doi.org/10.1029/2000RS002603.
Wood, R., 2005: Drizzle in stratiform boundary layer clouds. Part II: Microphysical aspects. J. Atmos. Sci., 62, 3034–3050, https://doi.org/10.1175/JAS3530.1.
Wood, R., C. S. Bretherton, D. Leon, A. D. Clarke, P. Zuidema, G. Allen, and H. Coe, 2011: An aircraft case study of the spatial transition from closed to open mesoscale cellular convection over the southeast Pacific. Atmos. Chem. Phys., 11, 2341–2370, https://doi.org/10.5194/acp-11-2341-2011.
Yamaguchi, T., G. Feingold, and J. Kazil, 2017: Stratocumulus to cumulus transition by drizzle. J. Adv. Model. Earth Syst., 9, 2333–2349, https://doi.org/10.1002/2017MS001104.
Zuidema, P., D. Leon, A. Pazmany, and M. Cadeddu, 2012a: Aircraft millimeter-wave passive sensing of cloud liquid water and water vapor during VOCALS-REx. Atmos. Chem. Phys., 12, 355–369, https://doi.org/10.5194/acp-12-355-2012.
Zuidema, P., and Coauthors, 2012b: On tradewind cumulus cold pools. J. Atmos. Sci., 69, 258–280, https://doi.org/10.1175/JAS-D-11-0143.1.
Zuidema, P., G. Torri, C. Muller, and A. Chandra, 2017: A survey of precipitation-induced atmospheric cold pools over oceans and their interactions with the larger-scale environment. Surv. Geophys., 38, 1283–1305, https://doi.org/10.1007/s10712-017-9447-x.
