Abstract

Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) radio occultation (RO) refractivity profiles in altostratus and nimbostratus clouds from 2007 to 2010 are first identified based on collocated CloudSat data. Vertical temperature profiles in these clouds are then retrieved from cloudy refractivity profiles. Contributions of cloud liquid water content and ice water content are also included in the retrieval algorithm. The temperature profiles and their lapse rates are compared with those from a standard GPS RO wet retrieval without including cloud effects. On average, the temperatures from cloudy retrieval are about 0.5–1.0 K warmer than the GPS RO wet retrieval, except for the altitudes near the nimbostratus base. The differences of temperature between the two methods are largest in summer and smallest in winter. The lapse rate in altostratus clouds is around 6.5°–7.5°C km−1 and does not vary greatly with height. On the contrary, the lapse rate increases significantly with height in nimbostratus clouds, from about 2.5°–3.5°C km−1 near the cloud base to about 5.0°–6.0°C km−1 at cloud center and 6.5°–7.5°C km−1 below the cloud top. Seasonal variability of lapse rate derived from the cloudy retrieval is larger than that derived from the wet retrieval. The lapse rate within clouds is smaller in summer and larger in winter. The mean lapse rate decreases with temperature in all seasons.

1. Introduction

The global positioning system (GPS) radio occultation (RO) limb-sounding technique makes use of radio signals from the GPS satellites for sounding the earth's atmosphere with high vertical resolution in all weather conditions (Ware et al. 1996; Rocken et al. 1997; Anthes et al. 2000). Each GPS RO event provides a vertical profile of atmospheric refractivity, which is a function of atmospheric temperature, water vapor, and pressure profiles in clear-sky conditions. In the presence of clouds, atmospheric refractivity is also a function of liquid water content and ice water content besides temperature, water vapor, and pressure variables (Kursinski et al. 1997; Zou et al. 2012). The Constellation Observing System for Meteorology, Ionosphere and Climate/Formosa Satellite Mission 3 (COSMIC/FORMOSAT-3: referred to as COSMIC for brevity) consists of six low earth-orbiting (LEO) satellites instead of one, providing thousands of daily GPS RO temperature profiles (Anthes et al. 2008; Fong et al. 2009). The GPS RO measurement precision for the raw measurements of excess phase delay L is less than 1 mm, which is much smaller than the atmospheric delays of about 1 km in the lower troposphere. However, the accuracy of the GPS RO refractivity profiles in the troposphere below ~5 km is about 1%, which comes from the effect of the horizontal variations of atmospheric refractivity on GPS RO signals being not considered in the Abel inversion, as well as errors associated with the length of recorded radio occultation signals, additive noise, and some tunable inversion parameters (Kursinski et al. 1995; Kursinski and Hajj 2001; Sokolovskiy et al. 2010).

GPS RO techniques provide a unique opportunity for obtaining and examining temperature profiles and lapse rate within clouds. Lin et al. (2010) developed a cloudy RO retrieval using cloud measurements from a 94-GHz, nadir-pointing cloud-profiling radar (CPR) onboard CloudSat. In this study, the same technique is used for deriving RO temperature profiles and lapse rate climatology within altostratus and nimbostratus clouds in high latitudes (e.g., poleward of 40° latitude) in both hemispheres.

Temperature profiles within clouds are important for validation and improvement of climate and weather forecast models (Stephens et al. 1990; Diak et al. 1998; Bayler et al. 2000). Previously, in situ dropsonde measurements from aircraft and weather balloons (e.g., radiosondes) have been the primary data sources for studying in-cloud thermodynamic structures. However, these in situ measurements are rare in high latitudes, especially near polar regions. Operational weather satellites provide abundant radiance measurements in high latitudes but cannot resolve vertical variations of the thermodynamic structure within clouds. GPS RO measurements complement very well the vast available radiance measurements by providing a much needed high vertical resolution.

Lin et al. (2010) proposed a cloudy retrieval algorithm to obtain vertical profiles of temperature within clouds. However, that study was based on a limited number of COSMIC cloudy RO profiles (about 200) collocated with CloudSat data within a 4-month period and with no cloud classification. This study applies the cloudy retrieval algorithm of Lin et al. (2010) to a 4-yr period (2007–10) and examines a seasonal dependence of cloud temperature and lapse rate climatology in altostratus and nimbostratus clouds derived from GPS RO data. The paper is arranged as follows: Section 2 provides a brief description of data during a four-year period from 2007 to 2010. Section 3 reviews a GPS RO cloudy retrieval algorithm that is employed for obtaining in-cloud vertical profiles of temperature from COSMIC RO refractivity data in altostratus and nimbostratus clouds located in high latitudes. The temperature profiles and lapse rate obtained by using the new cloudy retrieval algorithm are compared with those from the standard GPS RO wet retrieval and among four seasons. Conclusions and future work are provided in section 5.

2. A brief description of observations

a. GPS RO profiles

The COSMIC GPS RO global datasets of atmospheric parameters (vertical profiles of refractivity, temperature, water vapor pressure, etc.) have been made available to the user community by COSMIC Data Analysis and Archive Center (CDAAC) since 21 April 2006. GPS RO refractivity is derived from the raw RO measurements of the excess Doppler shift of the radio signals transmitted by GPS satellites, in which three-dimensional position and orbiting velocity of both GPS and LEO satellites are required and the refractive index in the atmosphere is assumed spherically symmetric. Useful information on data processing and quality assessment can be found in Sokolovskiy et al. (2009a,b) and Schreiner et al. (2010). The present study uses COSMIC soundings from 2007 to 2010.

GPS RO refractivity measurement represents an integrated refraction effect of the atmosphere along a ray path within a few hundred kilometers long and 1.5-km diameter tube that is centered at the perigee point (i.e., the point on the ray path that is closest to the Earth's surface). In other words, GPS RO measurements have a horizontal resolution of about 200 and 1.5 km in along-ray and cross-ray directions, respectively. The vertical resolution ranges from better than 100 m in the lower troposphere to approximately 0.5 km in the stratosphere. The RO temperature and water vapor pressure profiles from CDAAC are obtained using a one-dimensional variational data assimilation (1DVAR) approach, which is often called wet retrieval for brevity. The CDAAC wet retrieval employs the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis as the first guess field. Details of the 1DVAR system can be found in technical documentation (COSMIC Project Office 2005).

b. On choices of altostratus and nimbostratus clouds

By measuring the returned power backscattered by clouds, CPR provides measurements that are sensitive to cloud reflectivity, liquid water content (LWC) and ice water content (IWC). A great strength of the microwave radar instrument CPR is the ability to retrieve various microphysical parameters of clouds and precipitation, such as IWC, LWC, cloud-base and cloud-top heights of multiple cloud layers, and cloud types, from its measured radar reflectivity factor Z using retrieval algorithms (Stephens et al. 2002). CPR is a nadir-pointing instrument with a horizontal resolution [e.g., size of an effective CPR field of view (FOV)] of approximately 1.4 and 2.5 km in across- and along-track directions, respectively (Tanelli et al. 2008).

Different types of clouds have different macro- and microphysical properties and are usually produced by different dynamic processes (Hartmann et al. 1992; Chen et al. 2000). CloudSat CPR–measured cloudy profiles are classified into eight types: stratus (St), stratocumulus (Sc), cumulus (Cu), nimbostratus (Ns), altocumulus (Ac), altostratus (As), deep convective, and high cloud (Stephens et al. 2002). An appropriate retrieval algorithm is developed for a selected cloud type based on cloud temperatures from ECMWF reanalysis, upward radiance measurements from collocated Moderate Resolution Imaging Spectroradiometer (MODIS) measurements, cloud-base altitude, precipitation types (rain, snow, and drizzle), horizontal and vertical dimensions, and LWC. The permissible bounds for the maximum radar reflectivity factor in a particular vertical profile Zmax and the temperature at that level for each cloud type were established using an extended cloud dataset obtained over a 1-yr period from the Southern Great Plains Cloud and Radiation Testbed site (Wang and Sassen 2001). Other parameters characterizing cloud-type-dependent characteristics were obtained by combining space-based active CPR with collocated passive remote sensing data from MODIS. The altostratus clouds have a Zmax less than 10 dBZ and equaling −30 dBZ at −45°C, with the temperature at −30 dBZ ranging between −20° and −5°C, the highest Zmax frequency being at −10 dBZ and −25°C. A cloud base was located between 2 and 7 km. A horizontal dimension is about 1000 km, and vertical dimension is moderate. There is no rain and a nearly zero LWC within altostratus. The nimbostratus clouds are characterized by a Zmax ranging between −10 and 15 dBZ, a temperature at the vertical level of Zmax ranging between −25° and 10°C, the highest Zmax frequency located at about 5 dBZ and 0°C. A cloud base is below 4 km, and a cloud thickness is greater than 4 km. A horizontal dimension is about 1000 km and vertical dimension is thick.There is prolonged rain and snow and nonnegligible LWC.More details can be found in CloudSat Project (2001).

Because CloudSat CPR only measures the cloud profiles at its nadir at 1-km resolution, the cloud information over a 200-km segment of a GPS RO ray path centered at the perigee point is not available. A single CloudSat profile is likely not representative of cloud over a long segment of a GPS RO ray path near the perigee point. Therefore, a cloudy RO may contain segments of no cloud or other cloud types. To minimize the ambiguity introduced by the above-mentioned uncertainty on cloudy retrieval results, this study focuses on two stratus-type clouds, altostratus and nimbostratus, whose horizontal scales are around 1000 km and are larger than other types of clouds (Wang and Sassen 2001). Stratus clouds also have large scales but are not considered in this study because of their low altitudes (about 1–2 km) where GPS RO retrieval is not reliable, especially in low latitudes. CloudSat CPR also encounters challenges to detect the optically thin stratus, stratocumulus, and trade cumuli that are generally at 1–2-km altitudes (CloudSat Project 2001).

c. Collocation between GPS RO profiles and CloudSat CPR products

COSMIC RO measurements within clouds are identified from their collocations with CloudSat CPR data. Collocation between COSMIC GPS ROs and CloudSat CPR measurements is defined by a temporal difference of less than 0.5 h and a spatial separation of less than 30 km. Since the latitude and longitude of each GPS RO measurement for a single RO vertical profile change with altitude, a 30-km criterion of the spatial collocation is defined as the horizontal distance between the CPR cloud location and the COSMIC data location at the altitude of the cloud center of a collocated cloud profile. Under this collocation criterion, a total of 593 and 497 COSMIC profiles are found within altostratus and nimbostratus clouds, respectively, during a 4-yr period from 2007 to 2010. COSMIC GPS RO profiles in these two types of clouds within 40°N–40°S are eliminated from this study, which consists of only 84 and 10 collocated cases for altostratus and nimbostratus clouds, respectively. Geographical distributions of these selected COSMIC GPS RO cloudy profiles in altostratus and nimbostratus are provided in Fig. 1. Two of the reasons for more cloudy profiles in high latitudes are: (i) COSMIC GPS ROs are most populated near 50°N and 50°S [see Fig. 1a in Yang and Zou (2012)] and (ii) CloudSat CPR data are also more populated in higher latitudes.

Fig. 1.

Geographical distribution of GPS RO profiles collocated with altostratus (dots) and nimbostratus (plus signs) during a 4-yr period from 2007 to 2010. The collocation criteria are a temporal difference of less than 0.5 h and a spatial separation of less than 30 km.

Fig. 1.

Geographical distribution of GPS RO profiles collocated with altostratus (dots) and nimbostratus (plus signs) during a 4-yr period from 2007 to 2010. The collocation criteria are a temporal difference of less than 0.5 h and a spatial separation of less than 30 km.

Examples of CloudSat CPR–measured vertical profiles of IWC within altostratus and nimbostratus clouds are shown in Fig. 2. CloudSat IWC bias errors were estimated to be less than 40% (Austin et al. 2009). Each vertical profile is drawn from cloud base to cloud top for collocated cloudy ROs found between 61° and 50° S for altostratus clouds and between 82° and 68°S for nimbostratus clouds. It is seen that altostratus and nimbostratus clouds in high latitudes in the Southern Hemisphere are mostly ice clouds. Very little areas of IWC are found near the cloud base. The IWC in nimbostratus clouds is in general higher than in altostratus clouds. On average, nimbostratus clouds have a lower cloud base and are thicker clouds than altostratus clouds.

Fig. 2.

Vertical distribution of IWC (g m−3) within (a) altostratus clouds between 61° and 50°S and (b) nimbostratus clouds between 82° and 68°S plotted at their observed latitudes. Each vertical bar is drawn from cloud base to cloud top. Areas without IWC are indicated in black.

Fig. 2.

Vertical distribution of IWC (g m−3) within (a) altostratus clouds between 61° and 50°S and (b) nimbostratus clouds between 82° and 68°S plotted at their observed latitudes. Each vertical bar is drawn from cloud base to cloud top. Areas without IWC are indicated in black.

The horizontal distances between CloudSat and GPS RO data points of all collocated pairs are shown in Fig. 3, using either the mean latitudinal and longitudinal position of the GPS RO profile or the observing location of the collocated GPS RO event at the cloud center for both cloud types. It is seen that the distance between a pair of collocated CloudSat cloud and GPS RO can be significantly larger than 30 km if the mean position of a GPS RO profile is used for collocation calculation. In other words, many GPS RO profiles that are close to CloudSat observed clouds (30 km) would be eliminated if the mean position of a GPS RO profile were used for collocation calculation. Similarly, many GPS RO profiles that are far away (>30 km) from CloudSat observed clouds would be retained if the mean position of a GPS RO profile was used for collocation calculation (not shown).

Fig. 3.

Scatterplots of the horizontal distance between any CloudSat cloud and the mean location of its collocated GPS RO (y axis) and those between any CloudSat cloud and its collocated GPS RO location at the cloud center (x axis) for (left) altostratus and (right) nimbostratus clouds.

Fig. 3.

Scatterplots of the horizontal distance between any CloudSat cloud and the mean location of its collocated GPS RO (y axis) and those between any CloudSat cloud and its collocated GPS RO location at the cloud center (x axis) for (left) altostratus and (right) nimbostratus clouds.

Vertical profiles of temperature from the ECMWF analyses along the CloudSat observation granule at CloudSat data resolution are included in the CloudSat auxiliary data products (see CloudSat algorithm process description documents: available online at http://www.cloudsat.cira.colostate.edu/cloudsat_documentation/html/index.html). The relative humidity (RH) required by this study is from the ECMWF analyses. These ECMWF profiles are generated from the ECMWF analyses at the new T799 resolution, which corresponds to a 25-km horizontal resolution and 91 vertical levels with the model top located around 0.01 hPa.

3. GPS RO cloudy retrieval algorithm

In the neutral atmosphere, the GPS observed atmospheric refractivity is a function of the atmospheric pressure P, atmospheric temperature T, water vapor pressure Pw, LWC, and IWC through the following relationship:

 
formula

where P and Pw are both given in hectopascals, T is given in kelvin, and (the liquid water content) and (the ice water content) are both given in grams per cubic meter. The first term on the right-hand side of (1) is referred to as the “dry” term, the second is the “wet” term, the third is the liquid water content term, and the fourth is the ice water content term. Therefore, the atmospheric refractivity is a sum of contributions from dry atmosphere, water vapor, liquid water, and ice water. Zou et al. (2012) presented a theoretical derivation of the third and fourth terms in (1), accounting for the effects of liquid and ice clouds on the propagation of GPS radio signals. By averaging over a total of 3469 global GPS RO cloudy profiles during a 3-yr period from 2007 to 2009, Yang and Zou (2012) obtained the following results: the dry term contributes more than 90% to the total refractivity, the water vapor contributes a few percentages (8%–9%), and the liquid water and ice water contribute an additional 0.1%–0.3%.

The represents a weighted average of the atmospheric refraction [see (1)] along the ray path of a radio signal near the perigee point. Clouds may not occupy the ray path over which an integrated contribution of the atmospheric refraction is represented by a cloudy RO measurement or the ECMWF analysis grids. Therefore, similar to what is often done in numerical weather prediction (NWP) models in which convection is initiated when the relative humidity exceeds a threshold value (e.g., 85% instead of 100%), we introduce an empirical parameter in the retrieval of the RO temperature profiles within clouds. We may express an observed GPS refractivity as a weighted average of a clear-sky and a cloudy refractivity,

 
formula

where

 
formula
 
formula

where es is the saturation vapor pressure. The saturation assumption corresponds to a totally cloudy condition (i.e., ). The standard GPS RO wet retrieval is based on the clear-sky refractivity with a priori temperature and water vapor background information from the ECMWF reanalysis. It is also reminded that the last term was not included in Lin et al. (2010).

The parameter in (2) represents a mean RH along the GPS RO ray path (e.g., ~200 km) and is calculated based on ECMWF analysis. Figure 4 shows vertical distributions of the total number of collocations with altostratus and nimbostratus clouds with different values of IWC (Figs. 4a,b) and RH (Figs. 4c,d). Altostratus clouds are most populated between 5- and 7-km altitudes while nimbostratus clouds are throughout the low troposphere below about 6 km. The total number of data points with IWC smaller than 0.1 or 0.2 g m−3 is much higher than those with larger IWC values. The RH decreases with height, from a value of about 95% near the earth's surface to about 60% at about 6 km and 40% at 10 km. The mean RH calculated as a function of height is used for the parameter in (2), which is a function of height z,

 
formula

where z is in kilometers. In Lin et al. (2010), was expressed as a function of the vertically averaged IWC value provided by CloudSat instead of the height .

Fig. 4.

Vertical distributions of the total number of collocated cloud cases located within (a),(b) altostratus and (c),(d) nimbostratus clouds with different values of (top) IWC and (bottom) RH at different altitudes. The RH interval is 1%, and the IWC interval is 0.02 g m−3. The mean RH profiles are shown with black dots. The vertical profile of the parameter α(z) is shown by the thin black straight line.

Fig. 4.

Vertical distributions of the total number of collocated cloud cases located within (a),(b) altostratus and (c),(d) nimbostratus clouds with different values of (top) IWC and (bottom) RH at different altitudes. The RH interval is 1%, and the IWC interval is 0.02 g m−3. The mean RH profiles are shown with black dots. The vertical profile of the parameter α(z) is shown by the thin black straight line.

Based on (2), atmospheric refractivity in clouds is a function of temperature and pressure if qLWC and qIWC are provided from CloudSat CPR measurements. It is reminded that the averaged vertical profile of the ECWMF RH field is used for the parameter in (2) of . The cloudy retrieval derives a temperature profile for a given in-cloud profile of GPS RO refractivity (i.e., ), liquid water content (i.e., ), and ice water content (i.e., ), as well as an upper boundary conditions of temperature and pressure at the cloud top ( and , respectively). It starts from the cloud top and goes downward to the cloud base , where and are obtained from CloudSat. The and are defined as the RO wet temperature and pressure at . With known temperature and pressure at a specific vertical level (Tm and Pm at the mth level), the following hydrostatic equation is used for saturated cloudy air to derive the pressure at the (m + 1)th level, which is 0.1 km lower than the mth level:

 
formula

where g is the gravity constant with its altitude dependence accounted for. is the gas constant of dry air. Here, (6) is derived based on the ideal gas law and the hydrostatic equation assuming saturated water vapor pressure [see (2)–(5) in Lin et al. 2010].

At a specific level, once the pressure is obtained by a numerical finite-difference scheme applied (4), the in-cloud temperature for a collocated cloudy RO is obtained by finding the minimum of the following cost function:

 
formula

The value of is determined to be within the temperature range . The squared differences of refractivity between the GPS RO observations and model refractivity are calculated at different temperatures in this temperature range at an interval of 0.1°C. Variations of the squared differences of refractivity between Nobs and model simulations with temperature are provided in Fig. 5 for an arbitrarily chosen altostratus cloudy GPS RO that occurred at 1030 UTC 23 October 2008 with its mean position located at 51.02°S, 66.27°E. The altitude where the in-cloud temperature is solved is indicated in different colors. Note that the cloud center is located at 4.1-km altitude. It is seen that assumes an asymmetric parabolic function shape with larger gradients on the colder temperature side. The minimum of on each curve is taken as . The fact that lines along the x axis means that , implying that fitted very well in a least squares sense.

Fig. 5.

Variations of the squared differences of refractivity between Nobs and Nmodel for an arbitrarily chosen cloudy GPS RO that occurred at 1030 UTC 23 Oct 2008 with its mean position located at 51.02°S, 66.27°E. The altitude (km) at which Tcloud is to be found is indicated in color.

Fig. 5.

Variations of the squared differences of refractivity between Nobs and Nmodel for an arbitrarily chosen cloudy GPS RO that occurred at 1030 UTC 23 Oct 2008 with its mean position located at 51.02°S, 66.27°E. The altitude (km) at which Tcloud is to be found is indicated in color.

Cloud liquid water and ice water contents vary along the ray path. It is argued that the probability of having more cloudy data points with the identified cloud type is likely to be larger than the data points in any other category. Also, most contributions to GPS RO measurements come from the atmospheric state and cloudy conditions near the perigee point where the collocation determination is made. Therefore, statistical features of temperature profiles and lapse rates derived from GPS RO shall be representative of averaged thermodynamic states for altostratus and nimbostratus clouds. It is worth mentioning that altostratus clouds are horizontally homogeneous and nimbostratus clouds are not as homogeneous as altostratus clouds (Wang and Sassen 2001). When using the CloudSat IWC and LWC values in (1), the local spherical symmetric assumption employed in GPS RO retrieval of refractivity is more valid for altostratus than nimbostratus clouds.

4. Numerical results

a. Temperature profiles in clouds

The vertical profiles of Tcloud derived from minimizing differences between and in (2) are compared with Twet of COSMIC wet retrieval and observations in Figs. 6a,b for data points located at cloud centers. It is seen that the temperatures derived from the cloudy retrieval is in general warmer than the wet retrieval. A few cases are found for which Tcloud is colder than Twet and the saturation vapor pressure at the temperature Tcloud is smaller than the vapor pressure ewet obtained by the COSMIC wet retrieval. A schematic illustration of a case with Tcloud being greater than Twet and a case with Tcloud being smaller than Twet is provided in Fig. 7: the former occurred on 1030 UTC 23 October 2008 at 51.02°S, 66.27°E with its cloud center height of 4.0 km and the latter on 1336 UTC 11 December 2008 at 41.98°N, 165.85°W with its cloud center height of 3.4 km. Because of latent heat release, the temperature in clouds decreases at a smaller rate than in clear-sky environment. A wet retrieval without taking saturation conditions into account most probably produces a cold bias in clouds, which is shown in red in Figs. 6 and 7. The differences between Tcloud and Twet are larger at warmer temperatures. On the other hand, by taking the derivatives of the atmospheric refractivity with respect to temperature and water vapor, it is clear that changes of refractivity are negatively correlated to temperature increments and positively correlated to water vapor increments [see (1)]. Therefore, there may also be cases for which the water vapor from the wet retrieval is too low, which could result in a negative bias of the temperature in clouds.

Fig. 6.

Scatterplots for (a),(b) Tcloud vs Twet and (c),(d) variations of temperature differences (TcloudTwet) with respect to (ewet) at cloud centers of (left) altostratus and (right) nimbostratus. Data points with the differences between Tcloud and Twet (i.e., TcloudTwet) being positive and negative are indicated by red and blue dots, respectively.

Fig. 6.

Scatterplots for (a),(b) Tcloud vs Twet and (c),(d) variations of temperature differences (TcloudTwet) with respect to (ewet) at cloud centers of (left) altostratus and (right) nimbostratus. Data points with the differences between Tcloud and Twet (i.e., TcloudTwet) being positive and negative are indicated by red and blue dots, respectively.

Fig. 7.

A schematic illustration of an RO event with TcloudTwet ≥ 0 (red) and another RO event with TcloudTwet ≤ 0 (blue). The first RO event shown in red was located at 51.02°S, 66.27°E for 1030 UTC 23 Oct 2008 with the cloud center located at 4 km, and the second RO event was located at 41.98°N, 165.85°W for 1030 UTC 23 Oct 2008 with its cloud center located at 3.4 km.

Fig. 7.

A schematic illustration of an RO event with TcloudTwet ≥ 0 (red) and another RO event with TcloudTwet ≤ 0 (blue). The first RO event shown in red was located at 51.02°S, 66.27°E for 1030 UTC 23 Oct 2008 with the cloud center located at 4 km, and the second RO event was located at 41.98°N, 165.85°W for 1030 UTC 23 Oct 2008 with its cloud center located at 3.4 km.

Figure 8 presents the frequency distributions of differences of temperature (Figs. 8a,b) and water vapor pressure (Figs. 8c,d) between cloudy retrieval and GPS RO wet retrieval for all the data points within collocated clouds. The total number of data counts in different altitude ranges above cloud bases is also indicated. Temperatures and vapor pressures from the cloudy retrieval are systematically larger than those of the wet retrieval in both altostratus and nimbostratus clouds except for nimbostratus clouds near cloud bases.

Fig. 8.

Frequency distributions of differences of (a),(b) temperature and (c),(d) water vapor pressure between cloudy retrieval and GPS RO wet retrieval of (left) altostratus and (right) nimbostratus. The total number of data counts in different altitude ranges above cloud bases is indicated in color.

Fig. 8.

Frequency distributions of differences of (a),(b) temperature and (c),(d) water vapor pressure between cloudy retrieval and GPS RO wet retrieval of (left) altostratus and (right) nimbostratus. The total number of data counts in different altitude ranges above cloud bases is indicated in color.

The mean temperature differences are calculated by aligning all the clouds by cloud center (Figs. 9a,b), cloud top (Figs. 9c,d), and cloud base (Figs. 9e,f). It is found that the cloudy retrievals produced a warmer temperature throughout altostratus clouds with similar magnitudes. For nimbostratus clouds, the difference between Tcloud and Twet is largest at the cloud center and smallest near the cloud base. The number of cloudy cases with (TcloudTwet) 0 is significantly more than those cases with (TcloudTwet) < 0. The largest mean difference is about 1 K.

Fig. 9.

Mean of all data points (solid), only positive (TcloudTwet) (dotted), and all negative (TcloudTwet) (dashed) aligned at (a),(b) cloud center; (c),(d) cloud top; and (e),(f) cloud base for (left) altostratus and (right) nimbostratus. The data points with (TcloudTwet) 0 (<0) are indicated in gray (black) horizontal bars.

Fig. 9.

Mean of all data points (solid), only positive (TcloudTwet) (dotted), and all negative (TcloudTwet) (dashed) aligned at (a),(b) cloud center; (c),(d) cloud top; and (e),(f) cloud base for (left) altostratus and (right) nimbostratus. The data points with (TcloudTwet) 0 (<0) are indicated in gray (black) horizontal bars.

The seasonal variations of the temperatures in altostratus and nimbostratus clouds as well as their differences from the wet retrievals are shown in Fig. 10. As expected, temperatures in altostratus and nimbostratus clouds are highest in summer and lowest in winter (dashed curves in Figs. 10a–f). The differences between the cloudy retrieval and the wet retrieval are also largest in summer and smallest in winter (solid curves in Figs. 10a–e), except for summer nimbostratus clouds in the lower half of the clouds (red curve in Fig. 10f). The differences between Tcloud and Twet and the seasonal dependences of (TcloudTwet) are smallest in winter, fall, and spring within nimbostratus clouds below the cloud center.

Fig. 10.

Mean of temperature differences (TcloudTwet) (solid) and temperatures Tcloud (dashed) aligned at (a),(b) the cloud center; (c),(d) the cloud top; and (e),(f) the cloud base for (left) altostratus and (right) nimbostratus in four seasons.

Fig. 10.

Mean of temperature differences (TcloudTwet) (solid) and temperatures Tcloud (dashed) aligned at (a),(b) the cloud center; (c),(d) the cloud top; and (e),(f) the cloud base for (left) altostratus and (right) nimbostratus in four seasons.

b. Lapse rates

The lapse rates within clouds are shown in Figs. 11 and 12. It is seen that the mean lapse rate in both altostratus and nimbostratus clouds changes quite rapidly near the cloud base and cloud top. The mean lapse rate varies from 5.0° to 7.5°C km−1 (Figs. 11a,b,c). The variation of the mean lapse rate with height is smaller in altostratus clouds (Fig. 11) than in nimbostratus clouds (Fig. 12). The mean lapse rates in nimbostratus clouds are similar to those of altostratus above the cloud center but much smaller than those of altostratus near the cloud base, especially in summer (Figs. 11a, 12a). Also shown in Figs. 11 and 12 are lapse rates calculated from the wet retrieval Twet. The lapse rates for Twet are very close to those from Tcloud, except for nimbostratus and more so in the summer. The GPS wet retrieval produces a lapse rate profile that increases with height from about 3.0°C km−1 at the nimbostratus cloud base to about 6.0°C km−1 at the cloud center. The lapse rates of Tcloud are smaller than those of Twet near the nimbostratus cloud bases.

Fig. 11.

Mean lapse rates aligned at (a) the cloud center, (b) the cloud top, and (c) the cloud base for altostratus.

Fig. 11.

Mean lapse rates aligned at (a) the cloud center, (b) the cloud top, and (c) the cloud base for altostratus.

Fig. 12.

As in Fig. 11, but for nimbostratus.

Fig. 12.

As in Fig. 11, but for nimbostratus.

Figure 13 provides frequency distributions (Figs. 13a,b) and vertical distributions (Figs. 13c,d) of the frequency of lapse rates for altostratus and nimbostratus clouds. The frequency distributions of lapse rates peak between 6.0° and 8.5°C km−1 for both altostratus and nimbostratus clouds, with a larger spread for nimbostratus clouds than altostratus clouds. Altostratus clouds are more populated between 4 and 7 km, and nimbostratus clouds are more populated between 2 and 6.5 km.

Fig. 13.

(a),(b) Frequency distributions and (c),(d) vertical distributions of the frequency of lapse rates for (left) altostratus and (right) nimbostratus clouds within each of the 0.2°C km−1 lapse rate and 0.1-km-altitude intervals.

Fig. 13.

(a),(b) Frequency distributions and (c),(d) vertical distributions of the frequency of lapse rates for (left) altostratus and (right) nimbostratus clouds within each of the 0.2°C km−1 lapse rate and 0.1-km-altitude intervals.

The mean lapse rate decreases with temperature within both altostratus (Fig. 14a) and nimbostratus (Fig. 14b). The mean lapse rates in colder ice clouds are significantly higher than in warmer ice clouds. The standard deviation of the mean lapse rate is largest around −20°C for altostratus clouds and between −10° and −25°C for nimbostratus clouds (Figs. 14c,d).

Fig. 14.

Variations of the (top) mean and (bottom) std dev of lapse rate (°C km−1) with temperature within (a),(c) altostratus and (b),(d) nimbostratus.

Fig. 14.

Variations of the (top) mean and (bottom) std dev of lapse rate (°C km−1) with temperature within (a),(c) altostratus and (b),(d) nimbostratus.

5. Summary and conclusions

GPS RO provides vertical profiles of temperature and water vapor in cloudy regions over the globe with high vertical resolutions. In this study, we first derive vertical profiles of temperature and water vapor in altostratus and nimbostratus clouds located north of 40°N and south of 40°S from COSMIC GPS RO refractivity profiles, using a similar but not identical cloudy retrieval algorithm as Lin et al. (2010). Vertical structures of temperatures and lapse rates are then examined and compared with temperature profiles from the COSMIC wet retrieval. CloudSat data were used to determine the presence of clouds at the location of the RO observations.

By introducing a sounding-dependent relative humidity parameter α, we derived a set of in-cloud temperature profiles from a weighted sum of clear-sky and cloudy refractivity values. Relative to the standard wet retrieval, the cloudy temperature retrieval is consistently warmer within clouds, with the largest difference of ~1 K located at the cloud center level. The average lapse rate within clouds is nearly constant in altostratus clouds and the upper half of nimbostratus clouds.

Future work includes extending the present study to add data for 2011 and 2012 and to conduct independent validation to quantify the accuracy of the GPS RO retrieved cloud temperature using the proposed cloudy retrieval algorithm. Additionally, efforts will be made to explore the potential benefits of having high-vertical-resolution cloud temperature profiles provided by GPS RO techniques for improved simulation and assimilation of high-frequency (89–190 GHz) microwave cloudy radiances. A multiyear collocation among GPS ROs, CloudSat, and microwave radiance data from instruments onboard polar-orbiting meteorological satellites would allow this research to advance further. High-frequency microwave cloudy radiances measure both the emission and scattering, which are determined by cloud temperature and cloud droplet size distribution. Knowing the cloud temperature would allow a direct retrieval of hydrometeor parameters from cloudy radiance observations. The CloudSat data can be used for validation. This dataset with collocation among three types of observations could also be used for characterizing biases of cloudy radiances.

Acknowledgments

This research was jointly supported by Chinese Ministry of Science and Technology under 973 Project 2010CB951600 and the Priority Academic Program Development (PAPD) of Jiangsu Higher Education Institutions.

REFERENCES

REFERENCES
Anthes
,
R. A.
,
C.
Rocken
, and
Y.-H.
Kuo
,
2000
:
Applications of COSMIC to meteorology and climate
.
Terr. Atmos. Oceanic Sci.
,
11
,
115
156
.
Anthes
,
R. A.
, and
Coauthors
,
2008
:
The COSMIC/FORMOSAT-3 mission: Early results
.
Bull. Amer. Meteor. Soc.
,
89
,
313
333
.
Austin
,
R. T.
,
A. J.
Heymsfield
, and
G. L.
Stephens
,
2009
:
Retrieval of ice cloud microphysical parameters using the CloudSat millimeter-wave radar and temperature
.
J. Geophys. Res.
,
114
,
D00A23
,
doi:10.1029/2008JD010049
.
Bayler
,
G. M.
,
R. M.
Aune
, and
W. H.
Raymond
,
2000
:
NWP cloud initialization using GOES sounder data and improved modeling of nonprecipitating clouds
.
Mon. Wea. Rev.
,
128
,
3911
3920
.
Chen
,
T.
,
W. B.
Rossow
, and
Y. C.
Zhang
,
2000
:
Radiative effects of cloud-type variations
.
J. Climate
,
13
,
264
286
.
CloudSat Project
,
2001
: Level 2 cloud scenario classification product process description and interface control document. Jet Propulsion Laboratory Rep., 20 pp. [Available online at http://cloudsat.cira.colostate.edu/ICD/2B-CLDCLASS/Level2_Product_Doc_Cldclass.doc.]
COSMIC Project Office
,
2005
: Variational atmospheric retrieval scheme (VARS) for GPS radio occultation data. University Corporation for Atmospheric Research COSMIC Project Office Rep., 8 pp. [Available online at http://cosmic-io.cosmic.ucar.edu/cdaac/doc/documents/1dvar.pdf.]
Diak
,
G. R.
,
M. C.
Anderson
,
W. L.
Bland
,
J. M.
Norman
,
J. M.
Mecikalski
, and
R. M.
Aune
,
1998
:
Agricultural management decision aids driven by real-time satellite data
.
Bull. Amer. Meteor. Soc.
,
79
,
1345
1355
.
Fong
,
C.-J.
, and
Coauthors
,
2009
:
FORMOSAT-3/COSMIC Spacecraft Constellation System, Mission Results, and Prospect for Follow-On Mission
.
Terr. Atmos. Oceanic Sci.
,
20
,
1
19
,
doi:10.3319/TAO.2008.01.03.01(F3C)
.
Hartmann
,
D. L.
,
M. E.
Ockert-Bell
, and
M. L.
Michelsen
,
1992
:
The effect of cloud type on the earth's energy balance: Global analysis
.
J. Climate
,
5
,
1281
1304
.
Kursinski
,
E. R.
, and
G. A.
Hajj
,
2001
:
A comparison of water vapor derived from GPS occultations and global weather analyses
.
J. Geophys. Res.
,
106
,
1113
1138
.
Kursinski
,
E. R.
,
G. A.
Hajj
,
K. R.
Hardy
,
L. J.
Romans
, and
J. T.
Schofield
,
1995
:
Observing tropospheric water vapor by radio occultation using the global positioning system
.
Geophys. Res. Lett.
,
22
,
2365
2368
.
Kursinski
,
E. R.
,
G. A.
Hajj
,
J. T.
Schofield
,
R. P.
Linfield
, and
K. R.
Hardy
,
1997
:
Observing Earth's atmosphere with radio occultation measurements using the global positioning system
.
J. Geophys. Res.
,
102
,
23 429
23 465
.
Lin
,
L.
,
X.
Zou
,
R.
Anthes
, and
Y.
Kuo
,
2010
:
COSMIC GPS radio occultation temperature profiles in clouds
.
Mon. Wea. Rev.
,
138
,
1104
1118
.
Rocken
,
C.
, and
Coauthors
,
1997
:
Analysis and validation of GPS/MET data in the neutral atmosphere
.
J. Geophys. Res.
,
102
,
29 849
29 866
.
Schreiner
,
W.
,
C.
Rocken
,
S.
Sokolovskiy
, and
D.
Hunt
,
2010
:
Quality assessment of COSMIC/FORMOSAT-3 GPS radio occultation data derived from single- and double-difference atmospheric excess phase processing
.
GPS Solut.
,
14
,
13
22
,
doi:10.1007/s10291-009-0132-5
.
Sokolovskiy
,
S.
,
C.
Rocken
,
W.
Schreiner
,
D.
Hunt
, and
J.
Johnson
,
2009a
:
Postprocessing of L1 GPS radio occultation signals recorded in open-loop mode
.
Radio Sci.
,
44
,
RS2002
,
doi:10.1029/2008RS003907
.
Sokolovskiy
,
S.
,
W.
Schreiner
,
C.
Rocken
, and
D.
Hunt
,
2009b
:
Optimal noise filtering for the ionospheric correction of GPS radio occultation signals
.
J. Atmos. Oceanic Technol.
,
26
,
1398
1403
.
Sokolovskiy
,
S.
,
C.
Rocken
,
W.
Schreiner
, and
D.
Hunt
,
2010
:
On the uncertainty of radio occultation inversions in the lower troposphere
.
J. Geophys. Res.
,
115
,
D22111
,
doi:10.1029/2010JD014058
.
Stephens
,
G. L.
,
S. C.
Tsay
,
J. P. W.
Stackhouse
, and
P.
Flatau
,
1990
:
The relevance of the microphysical and radiative properties of cirrus clouds to climate and climatic feedback
.
J. Atmos. Sci.
,
47
,
1742
1753
.
Stephens
,
G. L.
, and
Coauthors
,
2002
: The CloudSat mission and the A-Train: A new dimension of space-based observations of clouds and precipitation. Bull. Amer. Meteor. Soc., 83, 1171–1790.
Tanelli
,
S.
,
S. L.
Durden
,
E.
Im
,
K. S.
Pak
,
D. G.
Reinke
,
P.
Partain
,
J. M.
Haynes
, and
R. T.
Marchand
,
2008
:
CloudSat's cloud profiling radar after two years in orbit: Performance, calibration, and processing
.
IEEE Trans. Geosci. Remote Sens.
,
46
,
3560
3573
.
Wang
,
Z.
, and
K.
Sassen
,
2001
:
Cloud type and macrophysical property retrieval using multiple remote sensors
.
J. Appl. Meteor.
,
40
,
1665
1682
.
Ware
,
R.
, and
Coauthors
,
1996
:
GPS sounding of the atmosphere from low Earth orbit: Preliminary results
.
Bull. Amer. Meteor. Soc.
,
77
,
19
40
.
Yang
,
S.
, and
X.
Zou
,
2012
:
Assessments of cloud liquid water contributions to GPS RO refractivity using measurements from COSMIC and CloudSat
.
J. Geophys. Res.
,
117
,
D06219
,
doi:10.1029/2011JD016452
.
Zou
,
X.
,
S.
Yang
, and
P. S.
Ray
,
2012
:
Impacts of ice clouds on GPS radio occulation measurements
.
J. Atmos. Sci.
,
69
,
3670
3682
.