Cloud Variability over the Indian Monsoon Region as Observed from Satellites

Margaret M. Wonsick Department of Atmospheric and Oceanic Science, University of Maryland, College Park, College Park, Maryland

Search for other papers by Margaret M. Wonsick in
Current site
Google Scholar
PubMed
Close
,
Rachel T. Pinker Department of Atmospheric and Oceanic Science, University of Maryland, College Park, College Park, Maryland

Search for other papers by Rachel T. Pinker in
Current site
Google Scholar
PubMed
Close
, and
Yves Govaerts European Organisation for the Exploitation of Meteorological Satellites, Darmstadt, Germany

Search for other papers by Yves Govaerts in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

This study focuses on documenting the seasonal progression of the Asian monsoon by analyzing clouds and convection in the pre-, peak-, and postmonsoon seasons. This effort was possible as a result of the movement of Meteosat-5 over the Indian continent during the Indian Ocean Experiment (INDOEX) starting in 1998. The Meteosat-5 observations provide a unique opportunity to study in detail the daytime diurnal variability of clouds and components of the radiation budget. Hourly Meteosat-5 observations are utilized to characterize the Indian monsoon daytime cloud variability on seasonal and diurnal time scales. Distinct patterns of variability can be identified during the various stages of the monsoon cycle. The daytime (0800–1500 LST) diurnal cycle of total cloud amounts is generally flat during the premonsoon season, U shaped during peak-monsoon season, and ascending toward an afternoon peak in the postmonsoon season. Low clouds dominate the Tibetan Plateau and northern Arabian Sea while high clouds are more frequent in the southern Bay of Bengal and Arabian Sea. An afternoon peak in high clouds is most prominent in central India and the Bay of Bengal. Afternoon convection peaks earlier over water than land. Preliminary comparison of cloud amounts from Meteosat-5, International Satellite Cloud Climatology Project (ISCCP) D1, and model output from the 40-yr ECMWF Re-Analysis (ERA-40) and the NCEP–NCAR reanalysis indicates a large disparity among cloud amounts from the various sources, primarily during the peak-monsoon period. The availability of the high spatial and temporal resolution of Meteosat-5 data is important for characterizing cloud variability in regions where clouds vary strongly in time and space and for the evaluation of numerical models known to have difficulties in predicting clouds correctly in this monsoon region. This study also has implications for findings on cloud variability from polar-orbiting satellites that might not correctly represent the daily average situation.

Corresponding author address: Margaret M. Wonsick, Department of Atmospheric and Oceanic Science, University of Maryland, College Park, College Park, MD 20742. Email: mwonsick@atmos.umd.edu

Abstract

This study focuses on documenting the seasonal progression of the Asian monsoon by analyzing clouds and convection in the pre-, peak-, and postmonsoon seasons. This effort was possible as a result of the movement of Meteosat-5 over the Indian continent during the Indian Ocean Experiment (INDOEX) starting in 1998. The Meteosat-5 observations provide a unique opportunity to study in detail the daytime diurnal variability of clouds and components of the radiation budget. Hourly Meteosat-5 observations are utilized to characterize the Indian monsoon daytime cloud variability on seasonal and diurnal time scales. Distinct patterns of variability can be identified during the various stages of the monsoon cycle. The daytime (0800–1500 LST) diurnal cycle of total cloud amounts is generally flat during the premonsoon season, U shaped during peak-monsoon season, and ascending toward an afternoon peak in the postmonsoon season. Low clouds dominate the Tibetan Plateau and northern Arabian Sea while high clouds are more frequent in the southern Bay of Bengal and Arabian Sea. An afternoon peak in high clouds is most prominent in central India and the Bay of Bengal. Afternoon convection peaks earlier over water than land. Preliminary comparison of cloud amounts from Meteosat-5, International Satellite Cloud Climatology Project (ISCCP) D1, and model output from the 40-yr ECMWF Re-Analysis (ERA-40) and the NCEP–NCAR reanalysis indicates a large disparity among cloud amounts from the various sources, primarily during the peak-monsoon period. The availability of the high spatial and temporal resolution of Meteosat-5 data is important for characterizing cloud variability in regions where clouds vary strongly in time and space and for the evaluation of numerical models known to have difficulties in predicting clouds correctly in this monsoon region. This study also has implications for findings on cloud variability from polar-orbiting satellites that might not correctly represent the daily average situation.

Corresponding author address: Margaret M. Wonsick, Department of Atmospheric and Oceanic Science, University of Maryland, College Park, College Park, MD 20742. Email: mwonsick@atmos.umd.edu

1. Introduction

Understanding monsoon variability is a primary goal of the Coordinated Energy and Water Cycle Observation Project (CEOP), a special initiative under the Global Energy and Water Cycle Experiment (GEWEX) of the World Climate Research Programme (WCRP). It emphasizes the need to document the advance and withdrawal of the monsoons as a basis for understanding their driving mechanisms (http://monsoon.t.u-tokyo.ac.jp/ceop/objectives.html). In alignment with this CEOP objective, this study focuses on documenting the seasonal progression of the Asian monsoon through an analysis of clouds and convection in the pre-, peak-, and postmonsoon seasons.

Satellite observations are a primary source of large-scale information on clouds. Of particular interest are observations from geostationary satellites that can depict the diurnal cycle of highly variable parameters that modulate the radiation budgets, such as clouds. Historically, information from geostationary satellites over the Indian subcontinent has not been as abundant as for other geographical locations. The International Satellite Cloud Climatology Project (ISCCP), a major program of WCRP for global-scale observations on clouds, is the most versatile source of such information (Schiffer and Rossow 1983, 1985; Rossow and Schiffer 1991, 1999; Rossow and Garder 1993a,b; Rossow et al. 1996; Brest and Rossow 1992). The data are suitable for inferring the global distribution of clouds, their properties, and their effect on the radiation balance. Data collection began on 1 July 1983 and continues presently. However, this database suffered for most of its duration from what is known as the “Indian Ocean Gap,” a region from about 50° to 95°E where only National Oceanic and Atmospheric Administration/Advanced Very High Resolution Radiometer (NOAA/AVHRR) morning and afternoon satellite observations are available. This greatly reduced the number of observations in the region and clustered the time of the observations to only one daytime observation per NOAA satellite. Although the region has been covered by the Indian National Satellite System (INSAT) series of geostationary satellites first launched in 1982, the entire record of INSAT data is not publicly available. The Indian Meteorological Department (IMD) has arranged to distribute a subset through the National Center for Atmospheric Research (NCAR) (Roca et al. 2005) for a 5-yr period with only about two frames per day. For this reason, the INSAT data are not incorporated into the ISCCP database. In support of the Indian Ocean Experiment (INDOEX) (Ramanathan et al. 2001), the European Meteosat-5 satellite was moved to 63°E longitude, and continuous operational coverage of the area began in July 1998. Meteosat-5 observations over India have been incorporated into the ISCCP dataset since 1998, but with spatial and temporal subsampling of 30 km and 3 h, respectively.

There are two other sources for geostationary coverage of the Indian monsoon region. Some investigators have used the Japanese Geostationary Meteorological Satellite (GMS), which is centered at 140°E longitude (Murakami 1983; Islam et al. 2004), although this is not the ideal data source as India is on the western limb of the satellite’s field of view. The current satellite of the INSAT series, KALPANA-1, was launched in 2002, and the data are beginning to be used in meteorological applications over India (e.g., Shyamala and Bhadram 2006).

Previous studies to depict the diurnal cycle of clouds and precipitation over the Indian monsoon region have used various sources of information. Yang and Slingo (2001) used the European Union Cloud Archive User Service (CLAUS) dataset to infer clouds from brightness temperature information in the 10.3–11.3-μm band. This dataset is based on the ISCCP B3 data sampled at 30 km and a temporal resolution of 3-hourly intervals. For the period of the study, only polar orbiter data are included in the ISCCP data over India, so the diurnal cycle is not well represented in this region. Roca and Ramanathan (2000) reported on the diurnal variation of convective clouds based on cloud size, as derived from 3-hourly INSAT data in the 10.5–12.5-μm band. Their study focused mainly on clouds during the winter monsoon season. Sorooshian et al. (2002) evaluated convection over the Bay of Bengal and Calcutta, India, from microwave and radar measurements of rainfall from the Tropical Rainfall Measuring Mission (TRMM) satellite. Janowiak et al. (1994) used passive microwave data from the Special Sensor Microwave Imager (SSM/I) instrument on polar-orbiting Defense Meteorological Satellite Program (DMSP) satellites to determine the diurnal cycle of rainfall. Diurnal characteristics were inferred by using several DMSP satellites with different nodal crossing times. Islam et al. (2004) tracked the diurnal cycle of clouds and precipitation in Bangladesh and a small part of the Bay of Bengal with IR brightness temperature data from the Japanese GMS-5 and radar data from the Bangladesh Meteorological Department. Sen Roy and Balling (2007) examined data from 78 land-based rain gauges over the Indian subcontinent for the summers of 1980–2000. A few studies have employed Meteosat-5 in limited contexts. Zuidema (2003) contrasted convection in the Bay of Bengal for two years based on IR brightness temperatures from INSAT in 1988 and Meteosat-5 in 1999. Krishnamurti and Kishtawal (2000) and Barros et al. (2004) used a combination of Meteosat-5 IR brightness temperature data along with rainfall observations from TRMM for studies of convection in the Asian monsoon region. The former study focused solely on break periods of the monsoon to detect a diurnal mode of convection, and did not address monsoon characteristics. The latter investigation concentrated on convection patterns in the Tibetan Plateau, Himalayas, and Ganges basin.

The focus of this study is to develop methodologies for deriving cloud properties over the Indian monsoon region, to show their applicability for detailed characterization of cloud variability during the monsoon season, and to demonstrate the usefulness of the data for evaluating similar parameters derived from numerical models. A new method is developed to derive high-resolution cloud parameters using the full (hourly) temporal resolution and 5-km spatial resolution of Meteosat-5 satellite observations. Specifically, a cloud mask algorithm (Li et al. 2007; Pinker et al. 2007) and a radiative flux inference scheme (Pinker et al. 2003) designed for Geostationary Operational Environmental Satellite (GOES)-8 observations over the United States have been adapted to process Meteosat-5 data over India. Derived parameters include total cloud amount, cloud-top temperature, and cloud optical depth. The latter two parameters are used to identify cloud layers and convection. Cloud characteristics are first analyzed using spatial cloud amount patterns averaged over the pre-, peak-, and postmonsoon seasons. The daylight diurnal cycle of clouds and convection are presented on an hourly time scale, stratified by different monsoon seasons. Since representation of clouds is recognized as a primary source of disagreement in numerical model predictions (IPCC 2007), total cloud amounts derived from Meteosat-5 and from ISCCP D1 are compared to cloud forecasts from the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40; Uppala et al. 2005) and from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis (Kistler et al. 2001) to gain insight on model similarities and differences. The years 2001 and 2003 are chosen for analysis. These are years with climatologically normal rainfall based on rain gauge observations as available from the Indian Institute of Tropical Meteorology (IITM) at Pune (Parthasarathy et al. 1995). Since the results are quite similar for both years, only the figures for 2001 are presented in most cases.

The current work adds to the results of previous studies of the Indian monsoon region in several ways. It offers a new cloud detection algorithm that incorporates visible satellite imagery along with brightness temperature information and exploits the high temporal resolution of Meteosat-5. It encompasses a larger area than some of the previous studies and contrasts the behavior of clouds in different parts of the monsoon region. Finally, it aggregates cloud observations in terms of premonsoon, peak-monsoon, and postmonsoon time frames to better understand the evolution of cloud patterns as the monsoon progresses.

2. Data

Hourly Meteosat-5 observations in the region bounded by 3°–39°N and 51°–95°E are the basis for the cloud analyses in this study. Data from the visible (0.4–1.0 μm) and infrared (10.5–12.5 μm) bands are used for clear/cloudy-sky determination. The data are collected at 2.5-km resolution in the visible band and 5-km resolution in the infrared band. The cloud detection algorithm is applied to pixels of 5-km resolution in both spectral bands and final results are projected onto a 0.125° latitude–longitude grid. The study covers the period of March–November 2001. The cloud detection algorithm requires knowledge of snow conditions to prevent misrepresentation of snow or ice as clouds. Snow cover information is obtained from the Interactive Multisensor Snow and Ice Mapping System (IMS) dataset, a manually developed global snow mask produced at NCEP (Ramsay 1998; NOAA/NESDIS/OSDPD/SSD 2008). Precipitable water is needed for the calculation of cloud optical depth, and such information is extracted from the NCEP reanalysis II, which is documented online (see http://www.cdc.noaa.gov/cdc/reanalysis/reanalysis.shtml.)

3. Methodology

In this study, total cloud amount and frequency of occurrence of low, high, and convective clouds in the Indian monsoon region are analyzed. Total cloud amount is derived from Meteosat-5 observations in the visible and IR channels, while cloud typing is based on IR brightness temperature and cloud optical depth. The methods used to derive such variables are explained below.

a. Cloud detection algorithm

The cloud detection algorithm used in this study is a simplified version of the Coupled Cloud and Snow Detection Algorithm (CCSDA) developed at the University of Maryland (Li et al. 2007; Pinker et al. 2007). The original algorithm employs two spatial variability tests and seven spectral threshold tests using combinations of four channels of GOES-8 data centered at 0.75, 3.9, 10.6, and 11.9 μm. The tests are adapted from the Clouds for AVHRR (CLAVR) cloud detection algorithm for the AVHRR sensor on the NOAA polar-orbiting satellites (Stowe et al. 1999), and concepts from the Moderate Resolution Imaging Spectroradiometer (MODIS) cloud mask (Frey et al. 2008; Ackerman et al. 2008) that was developed for the Terra and Aqua Earth Observing System (EOS) satellite missions. Table 1 shows that only three of these tests can be used for the Meteosat-5 cloud screening algorithm because of differences in the spectral resolution of the GOES-8 and Meteosat-5 sensors. Meteosat-5 provides observations in three spectral channels, centered at 0.75, 6.9, and 11.5 μm.

The three tests used for cloud detection are the Reflectance Gross Contrast Test (RGCT), the Thermal Gross Contrast Test (TGCT), and the Reflectance Uniformity Test (RUT). TCGT can be used for all hours, while RGCT and RUT can be used only during daylight hours when visible reflectance can be calculated. Reflectance is computed as follows:
i1558-8432-48-9-1803-e1
where L is narrowband radiance (W m−2 sr−1), E is the in-band extraterrestrial irradiance at mean sun–Earth distance (691 W m−2 sr−1 μm−1), δ is the square of the sun–Earth distance in astronomical units, and μ0 is the cosine of the solar zenith angle. The narrowband radiance L is given by
i1558-8432-48-9-1803-e2
where DC is the digital count from the satellite, DCo is the digital count offset for space, and Cf is the calibration coefficient that accounts for sensor degradation based on the time since the satellite was launched:
i1558-8432-48-9-1803-e3
where Cf 0 is the calibration coefficient at launch (0.8184 W m−2 sr−1/DC), Df is the daily drift (W m−2 sr−1/DC day−1 × 105), and DSL is the number of days since launch on 2 March 1991.

To monitor the daily drift, calibration of the visible sensor is performed about 4–8 times per year using simulations of radiative transfer for ocean and bright desert targets under multiple illumination conditions (Govaerts et al. 2004). The values of DCo and Df are updated with each calibration, and the history of their values can be found online (http://www.eumetsat.int/groups/ops/documents/document/pdf_ten_vis_calibdetail_met5.pdf.)

Cloud determination is made by aggregating 2 × 2 pixel squares and applying the three cloud tests referenced above. The RGCT compares the reflectance of each pixel to an empirically derived reflectance threshold set to 30 over ocean and 44 over land. Pixels with reflectances higher than the threshold values are classified as cloudy. Since snow-covered backgrounds have high reflectances that can be mistaken for clouds, this test is turned off when snow is present, as determined by the IMS snow dataset.

The TGCT uses IR brightness temperature thresholds and classifies pixels as cloudy if the IR brightness temperature sensed by the satellite is colder than the threshold value. The thresholds are currently set to 271 K over water and 249 K over land. This test works well as a supplement to the RGCT during daylight hours, but it is difficult to obtain a quality cloud analysis at night using only the TGCT. One of the objectives for developing a cloud detection methodology is to be able to use the cloud mask to derive solar radiation budgets and study the effect of aerosols on these budgets. Since such processes are of interest only during daylight hours, and since the powerful visible channel for cloud detection is missing during nighttime, it was opted to focus on cloud detection during daytime only. The monthly-mean cloud amounts presented in this study are daytime-only averages. A pixel is considered to be under daytime conditions when its solar zenith angle (the angle between the pixel’s zenith and the sun) is less than 75°.

The RUT examines the range of reflectances in the 2 × 2 pixel square and compares it to reflectance uniformity thresholds for cloudy and clear conditions. If the RGCT and TGCT tests detect clear conditions, but the difference in reflectances in the 2 × 2 square is more variable than what is expected under clear conditions, the pixels are flagged as mixed clear. Similarly, pixels identified as cloudy are tagged as mixed cloudy if the RUT difference is less variable than what is expected under cloudy conditions. The mixed pixels are then redistributed during the gridding process, as explained below.

Once the cloud classification has been performed at the pixel level, the data are reprojected onto a latitude/longitude grid. The gridding process requires the introduction of a clear-sky composite reflectance (CCR) to be used as a threshold in the reclassification of mixed pixels as either clear or cloudy. The CCR is computed for each pixel by averaging the reflectance values at each pixel for all clear-sky cases over the course of a month. A separate CCR is calculated for each hour of the day to account for variations in bidirectional reflectance. To make the final clear or cloudy determination for pixels flagged as mixed with the individual cloud tests, the reflectances of the mixed pixels in the grid box are compared to two values: the CCR and an empirically determined cloudy threshold. Mixed pixels with reflectance values less than or close to the CCR value are flagged as clear and mixed pixels with reflectance values larger than or close to the cloudy threshold are flagged as cloudy. The final cloud amount is calculated as the number of cloudy pixels in a grid box divided by the total number of pixels in the grid box. Other variables provided in the gridded projection include clear-sky reflectance, cloudy reflectance, snow amount, and CCR.

b. Cloud typing algorithm

Meteosat-5 IR brightness temperature is used to differentiate between low and high clouds. Low clouds are assumed to have cloud-top temperatures >263 K, while clouds with temperatures <250 K are classified as high clouds. This is similar to the high-cloud threshold of 255 K used by Roca et al. (2005), which was found to encompass the majority of convection-induced high cloudiness. The cloud classification procedure from MODIS uses multiple spectral bands, but case studies show that the 11–12-μm channel brightness temperatures for low, middle, and high clouds generally fall into the ranges given above (Li et al. 2003). To differentiate between upper-level cloudiness and convection, convective clouds must meet the high-cloud temperature threshold and have an optical depth ≥23. The optical depth criterion was chosen in accordance with the algorithm used by Rossow and Schiffer (1991) to identify convective clouds within the ISCCP data. Cloud optical depth information is estimated with the University of Maryland (UMD) Surface Radiation Budget (SRB) model as described by Pinker et al. (2003). The cloud information derived from the Meteosat-5 observations (section 3a) is used to drive the SRB model that produces radiative fluxes as well as estimates of cloud optical depth.

4. Cloud amount comparison with MODIS and ISCCP

Although there is no reliable “ground truth” for validation of cloud mask algorithms, comparison with other satellite-derived datasets has been widely used (Thomas et al. 2004; Hou et al. 1993; Rossow et al. 1993; and others). The various satellite products are influenced by spatial and temporal resolution, synchronization of observation times, and geographical collocation of grid points. However, they do provide useful information on the similarities and differences between cloud detection algorithms. Presented here is a comparison of clouds derived from Meteosat-5, MODIS, and ISCCP.

a. MODIS

The high spectral resolution of the MODIS instrument aboard the Terra and Aqua satellites launched in 1998 and 2002, respectively, has significantly enhanced satellite cloud detection capabilities. MODIS observes in 36 spectral bands between 0.415 and 14.235 μm with a spatial resolution of 250–1000 m depending on the spectral interval. The MODIS cloud detection algorithm uses a series of pixel-level tests comparing observed values in various spectral bands or combinations of bands to fixed threshold values (Frey et al. 2008; Ackerman et al. 2008). The most significant difference between the MODIS and Meteosat-5 cloud detection algorithms is the use of tests from more spectral bands with MODIS. In particular, the 1.38- and 7.7-μm bands are extremely useful for detecting thin cirrus clouds that do not have a strong visible signature and whose brightness temperature can appear lower due to contamination from the underlying warmer surface. These clouds can be missed using only the visible and IR bands of Meteosat-5.

A comparison of total cloud amount from Meteosat-5 and MODIS Collection 5 data over the Indian monsoon region is conducted for the months of April and July 2003. Temporal matching of observations is an issue because Meteosat-5 observes hourly while the Terra and Aqua satellites pass the region once per day with equator crossing times of 1030 and 1330 LST, respectively. To produce an approximate matchup, the Meteosat-5 observations are sorted to LST for each grid point and cloud amounts are averaged separately for 1000 and 1100 LST (comparable to MODIS Terra) and 1300 and 1400 LST (comparable to MODIS Aqua). The Meteosat-5 pixel-level cloud mask results are gridded at a spatial resolution of 1° × 1° to match that of the MODIS data.

Presented in Fig. 1 are monthly mean cloud amounts from Meteosat-5 and MODIS Terra averaged zonally over longitudes 51°–95°E for April 2003 (Fig. 1a) and July 2003 (Fig. 1b). Results from MODIS Aqua are very similar and therefore are not shown. There is a very good agreement between the cloud detection algorithms from 3° to about 25°N. Larger differences occur in the Tibetan Plateau where Meteosat-5 cloud amounts are up to 20%%–25% higher than MODIS. This region is characterized by short-lived convective clouds with scales of 1–2 km that move quickly because of the windy conditions of the plateau (Yeh and Gao 1979). Under these circumstances the cloud analyses are very sensitive to observation time, so the error associated with time interpolating the Meteosat-5 data to the MODIS observation time becomes more significant.

A tendency for the MODIS cloud detection algorithm to underestimate low clouds in the Tibetan Plateau was reported by Li et al. (2006), who compared a surface climatology of clouds over the Tibetan Plateau with clouds derived from MODIS and ISCCP. There is a natural bias in surface- versus satellite-observed low clouds because some low clouds will be obscured by high clouds from the satellite view. However, while ISCCP showed only a 20% underestimation of low cloud amount over the Tibetan Plateau, MODIS had a cloud frequency bias of up to 50%. Comparing Meteosat-5 versus MODIS, low cloud amounts separately from total cloud amount, it is found that Meteosat-5 low cloud amounts over the Tibetan Plateau are, on average, 8% higher than MODIS. Therefore, the Meteosat-5 observations appear to produce a better representation of low clouds over the Tibetan Plateau than MODIS.

b. ISCCP

The ISCCP cloud detection algorithm is described extensively in Rossow and Garder (1993a). ISCCP satellite inputs over the Indian monsoon domain for the period of this study include Meteosat-5 and two NOAA polar-orbiting satellites. The Meteosat-5 data are subsampled in time every 3 h (0000, 0300, 0600, 0900, 1200, 1500, 1800, 2100 UTC) and the NOAA satellites view the region twice per day (one daytime and one nighttime observation). The native spatial resolution of the satellite data is about 5 km but cloud detection is applied to pixels that are subsampled and mapped to a 30-km grid under the assumption that the radiance values of the selected pixels are representative of conditions for the 30-km grid box. While this assumption can be valid for large-scale cloud systems, it can lead to an under- or overestimate of cloud amount for small-scale broken cloud fields depending on whether the sampled pixel is clear or cloudy.

The Meteosat-5 and ISCCP cloud detection algorithms are similar in that both use spatial variability and threshold contrast tests in the visible and IR channels, but their implementation is different. ISCCP first applies a spatial variability test over small domains of 90 km × 90 km over land and ice and 450 km × 450 km over water. Pixels with IR brightness temperatures much colder than the maximum IR brightness temperature in the surrounding domain are flagged as cloudy by the spatial variability test. If these pixels also show significant IR brightness temperature variability over three consecutive days, they are classified as cloudy. Additional cloud tests are applied using clear-sky statistics of IR brightness temperature and visible reflectance that are compiled once every 5 days. Pixels that are significantly colder or brighter than the expected clear-sky background are classified as cloudy. Cloud fraction is computed for 280-km equal-area grid boxes (∼2.5° × 2.5° resolution at equator) by dividing the number of cloudy pixels by the total number of pixels in the grid box, with the assumption that pixels are either completely clear or completely cloudy.

A comparison of total cloud amount from ISCCP and Meteosat-5 over the Indian monsoon region is conducted for the months of April and July 2003. The hours of 0300, 0600, and 0900 UTC are averaged for each dataset to produce daytime mean cloud amounts that are subsequently averaged for each month. The Meteosat-5 cloud mask algorithm is applied at pixel level but the results are gridded to a spatial resolution of 2.5° × 2.5° to match that of the ISCCP data.

Presented in Fig. 2 are monthly mean cloud amounts from Meteosat-5 and ISCCP averaged zonally over longitudes 51°–95°E for April 2003 (Fig. 2a) and July 2003 (Fig. 2b). In April the Meteosat-5 cloud amounts are 10%–20% lower than ISCCP in the majority of the domain. In July there is closer agreement except in latitudes higher than about 27°N. The most likely sources of differences are the cloud detection thresholds and the differing resolutions at which the cloud detection tests are applied. The most difficult cloud type for the Meteosat-5 algorithm to detect with its visible and IR contrast tests is thin cirrus. It is possible that the temporal variability test of ISCCP provides an advantage in detection of this cloud type, which in turn contributes to the higher cloud amount estimates of ISCCP in April. However, the MODIS 1.38-μm channel is extremely effective for detecting thin cirrus, and there is not a similar discrepancy in cloud amount between Meteosat-5 and MODIS in April (Fig. 1a). The effect of spatial resolution on satellite cloud detection was addressed by Wielicki and Parker (1992). They studied the performance of several cloud detection techniques at varying resolutions and found that applying the ISCCP algorithm to broken cumulus cloud fields resulted in a 5% overestimation of cloud amount in comparison with a reference cloud amount determined from high-resolution Landsat imagery. The spatial resolution may partially explain the larger differences between Meteosat-5 and ISCCP cloud amounts during April, when broken cumulus fields are more prevalent versus the widespread cloudiness over the Indian subcontinent that occurs during the peak of the monsoon in July. Unlike in the MODIS comparison, cloud amounts in the Tibetan Plateau in July are fairly similar between ISCCP and Meteosat-5 with differences no larger than 10%. The higher differences in zonal mean cloud amounts in latitudes 27°–40°N in July (Fig. 2b) are attributable to lower Meteosat-5 cloud amounts west of the plateau in Afghanistan and Iran.

5. Results

a. Spatial patterns of cloud amount

To document the progression of the monsoon, the mean total daytime cloud amount for the 2001 and 2003 premonsoon (March–May), peak-monsoon (June–September), and postmonsoon (October–November) seasons is computed and presented in Fig. 3. In the premonsoon phase (Fig. 3, top panels), the Indian subcontinent and surrounding oceans experience no more than 30% cloudiness. This pattern changes radically during the peak-monsoon months (Fig. 3, middle panels), when mean cloud amounts increase to 50%–80% in northeastern India and the Arabian Sea, and up to 90% over a large portion of the northern Bay of Bengal. Northwestern India, however, does not experience a large increase in cloud cover. This may be explained by the fact that the monsoon wind flow is northwestward from the Bay of Bengal onto the subcontinent, and most of the moisture is depleted before reaching the northwestern region. An alternative explanation is that the region remains relatively cloud free because the northward extent of the migration of the ITCZ is about 25°N. Postmonsoon cloud amounts (Fig. 3, bottom panels) drop dramatically throughout most of the subcontinent. In 2001 it is particularly evident that cloudiness continues during this phase in the eastern portion of the southern peninsula, which typically receives most of its rainfall during the months of October and November (Gadgil 2003). Regions of high cloud amount over water progress southward as the monsoon subsides.

Figure 4 shows the change in monthly mean total daytime cloud amount as the monsoon moves through different phases for the 2001 season (2003 is not shown as the results are substantially the same). During the monsoon buildup from May into June (Fig. 4a), the maximum increase in cloudiness occurs in latitudes 10°–20°N. At the height of the monsoon (July into August; Fig. 4b), the zone of maximum increase in cloudiness pushes to 25°N, and then drops to 5°–15°N during the monsoon dissipation phase (September into October; Fig. 4c).

b. Daytime diurnal cycle of cloud amount

The amplitude of the daytime diurnal cycle of low clouds and high clouds is shown in Fig. 5. The amplitude is defined as the difference between the maximum and minimum percentage of cloud amount for daylight hours for each day, averaged over the premonsoon, peak-monsoon, and postmonsoon seasons of 2003. The low- and high-cloud classifications are based on cloud-top temperature as seen from the satellite. As such, a high-cloud classification does not preclude the presence of low clouds underneath, or a low-cloud base as in the case of deep convective clouds.

In the premonsoon season (Figs. 5a,b) cloud fraction is low throughout the region and there is very little diurnal variation in low-cloud amount (Fig. 5a). The exception is just north of the Bay of Bengal, where moisture from southerly flow over the bay condenses when it encounters the Himalaya range, leading to a stronger diurnal cycle. There is more diurnal variability in high clouds over the southern peninsula and the Tibetan Plateau (Fig. 5b). Over oceans, the amplitude of the diurnal cycle of high clouds is much smaller, in agreement with Bergman and Salby (1996), who found a nearly uniform diurnal cycle of tropical oceanic convection.

In the peak-monsoon season (Figs. 5c,d) low-cloud diurnal variability increases strongly over land with the northward push of monsoon clouds (Fig. 5c). Despite the increase of convective clouds in the southern Arabian Sea with the northward progression of the monsoon, diurnal variability increases only slightly, as observed by Rozendaal et al. (1995) who showed that only nonconvective marine low-level clouds exhibit large diurnal variability. High cloud amount becomes more variable in northern India but is reduced sharply over the Tibetan Plateau (Fig. 5d). In the postmonsoon season (Figs. 5e,f), diurnal variability is still high in the eastern part of the southern peninsula, which receives most of its annual rainfall in October, but the retreat of the monsoon is evident with a reduction in the amplitude of the diurnal cycle of clouds in the northern parts of the subcontinent.

To illustrate the variability of the diurnal cycle in different parts of the Indian monsoon region with an emphasis on the times at which clouds most frequently occur, six points have been chosen for analysis (Fig. 6). The points are chosen based on the location of the largest changes in cloud amount between the different monsoon seasons, as shown in Fig. 4. Figures 7a–c give the mean total cloud amount at these points averaged over the 2001 premonsoon, peak-monsoon, and postmonsoon season for the hours 0800–1500 LST, respectively. In the premonsoon season, the diurnal cycle is rather flat at most locations, although the northern Arabian Sea and both points in the Bay of Bengal display a morning maximum of clouds. In the peak-monsoon season, all points show a U-shape distribution, with diminishing cloud amounts in the middle of the day. A pronounced afternoon buildup of clouds is observed in the Tibetan Plateau. Afternoon cloudiness dominates at nearly every location in the postmonsoon season.

The frequency distribution of occurrence of low clouds (Fig. 8, lhs) and high clouds (rhs) is given for the 2001 premonsoon (top panels), peak-monsoon (middle panels), and postmonsoon (bottom panels) seasons. The low- and high-cloud classifications are based on cloud-top temperature as seen from the satellite. As such, a high-cloud classification does not preclude the presence of low clouds underneath, or a low-cloud base as in the case of deep convective clouds. There are several notable features in the distributions. The northern Arabian Sea is dominated by low clouds in the peak-monsoon season, indicating that although the northward progression of the monsoon brings an increase in cloud cover to the region, it is not of a convective nature. These clouds reach their maximum in the morning. Alternatively, the southern Arabian Sea has more high than low clouds during both the pre- and peak-monsoon seasons, and there is little diurnal variability for either type. Cloud behavior in the southern Bay of Bengal is quite similar to the southern Arabian Sea. For all seasons, the Tibetan Plateau is dominated by low clouds that peak in the afternoon. In central India and the northern Bay of Bengal, the diurnal cycle of low clouds is relatively flat for the pre- and peak-monsoon seasons, with a slight increase in afternoon low clouds during the postmonsoon season. The only noticeable diurnal signal of high clouds at these points occurs with the afternoon buildup of convective clouds during the peak of the monsoon.

c. Daytime diurnal cycle of convection

Since the majority of the monsoon rainfall is linked to convection, it is of interest to analyze convective clouds separately from the total cloud amounts. In Fig. 9, the frequency of occurrence of convection for selected daytime hours is shown throughout the Indian monsoon region during the peak-monsoon season of 2001. The daytime diurnal maximum of convection occurs in the afternoon over both land and water, although the peak is earlier (1300 LST) at ocean points. The diurnal variation over the Bay of Bengal is consistent with the patterns reported in Zuidema’s (2003) study of convection over the bay during the 1998 and 1999 monsoon seasons. Many studies of the tropics as a whole agree that convection over land reaches a maximum in late afternoon (Gray and Jacobson 1977; McGarry and Reed 1978; Murakami 1983; Janowiak et al. 1994; Dai 2001; Gambheer and Bhat 2001; Nesbitt and Zipser 2003; Islam et al. 2004). The time of maximum frequency of convection at the selected point in central India matches the value derived by Sen Roy and Balling (2007) using land-based rain gauges. An early morning maximum in tropical ocean convection has been reported by Sui et al. (1997), Yang and Slingo (2001), and Dai (2001). This is attributed to higher relative humidities and the nocturnal nature of mesoscale convective systems. Since the Meteosat-5 observations analyzed in this study do not cover the nondaylight hours, this diurnal feature is neither confirmed nor contradicted. However, other studies do agree with the observed afternoon convective peak over tropical oceans seen here. Both ship and satellite data employed by McGarry and Reed (1978) in their study of the Global Atmospheric Research Program (GARP) Atlantic Tropical Experiment (GATE) region yielded a convective cloudiness/rainfall maximum at 1400 LST, seemingly associated with large-scale convergence. Sorooshian et al. (2002) analyzed TRMM data in the Bay of Bengal and observed an extended period of convective rainfall that began in early morning hours and reached its peak at 1400 LST, while Islam et al. (2004) found the maximum in the bay to occur between 1400 and 1600 LST. Gambheer and Bhat (2001) reported a peak in formation of new convective systems in the waters surrounding India between 1500 and 0000 LST. Somewhat east of the Indian monsoon region at the Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) observation site, Sui et al. (1997) noted that the deepest convection displayed an early morning ocean maximum, but warmer convective clouds were most frequently observed in the afternoon due to the diurnal cycle of ocean skin temperature. These results were corroborated by Janowiak et al. (1994), who found that clouds with IR brightness temperatures > 235 K appeared mainly in the afternoon. Figure 10 shows the mean temperature of convection averaged over the 2001 peak-monsoon season (for selected daytime hours). The main feature to note is that the coldest convection occurs in the earlier hours over ocean and in the later hours over land. This is in line with reports of the strongest ocean convection occurring in the early morning hours (not included in this study) and the conclusion that the secondary afternoon peak in ocean convection occurs within warmer cells. The consistency of the present convective cloud analysis with previous findings is encouraging. The fact that this dataset has higher temporal and spatial resolution than previous products opens up a wealth of possibilities for applications of the data. One such application is the validation of numerical models. A preliminary evaluation is described in the next subsection.

d. Comparison with numerical models

To demonstrate the utility of this newly available cloud information for numerical model evaluation, total cloud amounts derived from Meteosat-5 are compared with cloud amounts from two model sources: ERA-40, a reanalysis of the numerical weather prediction model run by ECMWF, and the NCEP–NCAR reanalysis, a reanalysis of NCEP model data enhanced by assimilating other observational data sources. Data from 0600 UTC were chosen for the comparison, as the model data are available in 6-hourly increments (0000, 0600, 1200, and 1800 UTC) daily, and 0600 UTC is a daylight hour for the entire Meteosat-5 domain in this study. Figure 11 shows an instantaneous picture of total clouds at 0600 UTC 1 July 2001 from Meteosat-5 (Fig. 11a) and ERA-40 (Fig. 11b), as well as the actual IR and visible satellite imagery for the corresponding time (Figs. 11c,d, respectively). The Meteosat-5 cloud amounts are gridded at 0.125° resolution. ERA-40 data are also shown at 0.125° resolution, but smoothing is apparent since it was interpolated from the ECMWF model output on an N80 reduced Gaussian grid with approximate resolution of 1.125°. The NCEP data are not shown because of the much larger disparity in resolution, which is ∼2° for the T62 Gaussian grid. ERA-40 shows a pronounced overanalysis of clouds in comparison with the satellite-derived clouds, which match the actual satellite imagery quite well.

To assess the agreement of the datasets as a function of geographical location throughout the evolution of the monsoon, the monthly mean of the 0600 UTC total cloud amounts at the six points in Fig. 6 is calculated for each month of the 2001 monsoon season, and results are shown in Fig. 12. The ISCCP D1 data are included in the comparison to give a measure of the effect of space and time subsampling on the cloud analysis. (Further comparison of Meteosat-5 and ISCCP is given in Table 2, which shows daytime daily average cloud amounts from both sources.) To minimize possible concerns that spatial resolution of the products may affect the analysis, the Meteosat-5 data are averaged over 2.5° × 2.5° grid boxes and the 2.5° version of ERA-40 is used to match the ISCCP D1 resolution. The NCEP–NCAR reanalysis resolution is comparable at ∼2°. Of the four data sources, ERA-40 consistently produces the highest cloud amounts at all points throughout the season. On average, the ISCCP D1 cloud amounts are the next highest, followed by the NCEP–NCAR reanalysis. Meteosat-5 cloud amounts are generally lower than the other sources.

Chevallier and Kelly (2002) found that the spatial correlation between clouds derived from ECMWF and Meteosat-5 was lower over land than ocean. In this evaluation, agreement between all cloud sources is best over the northern parts of the Arabian Sea and Bay of Bengal, while larger variation exists in the southern parts of the water regions. The largest disparity in cloud amounts occurs for the Tibetan Plateau, where snow and highly variable terrain increase the difficulty of analysis, forecasting, and satellite retrieval of clouds. Agreement between the datasets seems to be greatest at the beginning and end of the monsoon season, with more variability during the peak-monsoon months. The lack of agreement in cloud forecasts shows evidence of the well-documented shortcomings of models in predicting Asian monsoon rainfall (Sperber and Palmer 1996; Krishnamurthy and Shukla 2000; Cherchi and Navarra 2003).

6. Summary and future work

Continuous geostationary satellite coverage over India is available from Meteosat-5 for almost 10 years, making it feasible to produce long-term cloud datasets for the region with high spatial and temporal resolution (observations in the region continue with Meteosat-7). This is a vast improvement over past satellite datasets, which do not completely cover the region, do not resolve the diurnal cycle, or are not readily available. This study presents a new algorithm for cloud detection from Meteosat-5 observations. Unlike much of the past work in this region, the cloud information in this study was aggregated for the pre-, peak-, and postmonsoon seasons to analyze the progression of the summer monsoon throughout its cycle. The hourly nature of the data was exploited to develop detailed characteristics of the diurnal cycle of total, low-, high-, and convective cloud amounts throughout the different phases of the monsoon and at different geographical locations. The uniqueness of cloud distributions for different seasons and locations became very obvious. For example, the daytime (8–15 LST) diurnal cycle of total cloud amounts was generally flat during the premonsoon season, U shaped during peak-monsoon season, and ascending toward an afternoon peak in the postmonsoon season. It was further shown that low clouds dominated the Tibetan Plateau and northern Arabian Sea while high clouds were more frequently seen in the southern Bay of Bengal and Arabian Sea. An afternoon peak in high clouds was most prominent in central India and the Bay of Bengal. The afternoon peak in convection occurred earlier over water than over land.

These observations have a potential for evaluation of climate model simulations of the monsoon. A comparison of total cloud amount derived from Meteosat-5, ISCCP D1, ERA-40, and the NCEP–NCAR reanalysis identified stronger differences during the peak-monsoon season than during other months. The high spatial resolution of the Meteosat-5 cloud data makes them also ideal for the evaluation of smaller-scale hydrological models. Future plans include the development of a nighttime cloud detection algorithm to enhance the usefulness of the dataset for model evaluation over the complete diurnal cycle.

In addition to cloud properties, the importance of radiative and microphysical aspects of the monsoon system is now gaining attention. The role of aerosols has been investigated by Chung et al. (2002), Babu et al. (2004), Patra et al. (2005), Lau et al. (2005), and many others. Future work will extend this analysis to surface radiation budget parameters. In particular, the interdependence of clouds, radiation, and aerosols will be addressed with the use of a unique aerosol climatology synthesized from satellite observations, model simulations, and ground measurements (Liu et al. 2005, 2008; Liu and Pinker 2008) that was recently incorporated into the UMD SRB radiative flux inference scheme.

Acknowledgments

This work was supported under NASA Grant NNG05GB35G to the University of Maryland. The authors thank the staff at the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) Archive and Retrieval Facility for providing the Meteosat-5 observations and consulting on the calibration. Thanks are also given to the other data providers: the NASA Langley Research Center Atmospheric Sciences Data Center, providers of the ISCCP data; the National Snow and Ice Data Center (NSIDC) at the University of Colorado, providers of the IMS snow data; the NOAA/ESRL CIRES Climate Diagnostics Center, providers of the NCEP–NCAR reanalysis data; and the NCAR ERA-40 archive, providers of the ERA-40 data.

REFERENCES

  • Ackerman, S. A., R. E. Holz, R. Frey, E. W. Eloranta, B. C. Maddux, and M. McGill, 2008: Cloud detection with MODIS. Part II: Validation. J. Atmos. Oceanic Technol., 25 , 10731086.

    • Search Google Scholar
    • Export Citation
  • Babu, S. S., K. K. Moorthy, and S. K. Satheesh, 2004: Aerosol black carbon over Arabian Sea during intermonsoon and summer monsoon seasons. Geophys. Res. Lett., 31 , L06104. doi:10.1029/2003GL018716.

    • Search Google Scholar
    • Export Citation
  • Barros, A. P., G. Kim, E. Williams, and S. W. Nesbitt, 2004: Probing orographic controls in the Himalayas during the monsoon using satellite imagery. Nat. Hazards Earth Syst. Sci., 4 , 2951.

    • Search Google Scholar
    • Export Citation
  • Bergman, J. W., and M. L. Salby, 1996: Diurnal variations of cloud cover and their relationship to climatological conditions. J. Climate, 9 , 28022820.

    • Search Google Scholar
    • Export Citation
  • Brest, C. L., and W. B. Rossow, 1992: Radiometric calibration and monitoring of NOAA AVHRR data for ISCCP. Int. J. Remote Sens., 13 , 235273.

    • Search Google Scholar
    • Export Citation
  • Cherchi, A., and A. Navarra, 2003: Reproducibility and predictability of Asian summer monsoon in the ECHAM4-GCM. Climate Dyn., 20 , 365379.

    • Search Google Scholar
    • Export Citation
  • Chevallier, F., and G. Kelly, 2002: Model clouds as seen from space: Comparison with geostationary imagery in the 11-μm window channel. Mon. Wea. Rev., 130 , 712722.

    • Search Google Scholar
    • Export Citation
  • Chung, C., V. Ramanathan, and J. T. Kiehl, 2002: Effects of the South Asian absorbing haze on the northeast monsoon and surface–air heat exchange. J. Climate, 15 , 24622476.

    • Search Google Scholar
    • Export Citation
  • Dai, A., 2001: Global precipitation and thunderstorm frequencies. Part II: Diurnal variations. J. Climate, 14 , 11121128.

  • Frey, R. A., S. A. Ackerman, Y. Liu, K. I. Strabala, H. Zhang, J. R. Key, and X. Wang, 2008: Cloud detection with MODIS. Part I: Improvements in the MODIS cloud mask for collection 5. J. Atmos. Oceanic Technol., 25 , 10571072.

    • Search Google Scholar
    • Export Citation
  • Gadgil, S., 2003: The Indian Monsoon and its variability. Annu. Rev. Earth Planet. Sci., 31 , 429467.

  • Gambheer, A. V., and G. S. Bhat, 2001: Diurnal variation of deep cloud systems over the Indian region using INSAT-1B pixel data. Meteor. Atmos. Phys., 78 , 215226.

    • Search Google Scholar
    • Export Citation
  • Govaerts, Y. M., M. Clerici, and N. Clerbaux, 2004: Operational calibration of the Meteosat radiometer VIS band. IEEE Trans. Geosci. Remote Sens., 42 , 19001914.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., and R. W. Jacobson, 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105 , 11711188.

  • Hou, Y. T., K. A. Campana, K. E. Mitchell, S. K. Yang, and L. L. Stowe, 1993: Comparison of an experimental NOAA AVHRR cloud dataset with other observed and forecast cloud datasets. J. Atmos. Oceanic Technol., 10 , 833849.

    • Search Google Scholar
    • Export Citation
  • Islam, M. N., T. Hayashi, H. Uyeda, T. Terao, and K. Kikuchi, 2004: Diurnal variations of cloud activity in Bangladesh and north of the Bay of Bengal in 2000. Remote Sens. Environ., 90 , 378388.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., P. A. Arkin, and M. Morrissey, 1994: An examination of the diurnal cycle in oceanic tropical rainfall using satellite and in situ data. Mon. Wea. Rev., 122 , 22962311.

    • Search Google Scholar
    • Export Citation
  • Kistler, R., and Coauthors, 2001: The NCEP–NCAR 50-Rear Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247267.

    • Search Google Scholar
    • Export Citation
  • Krishnamurthy, V., and J. Shukla, 2000: Intraseasonal and interannual variability of rainfall over India. J. Climate, 13 , 43664377.

  • Krishnamurti, T. N., and C. M. Kishtawal, 2000: A pronounced continental-scale diurnal mode of the Asian summer monsoon. Mon. Wea. Rev., 128 , 462473.

    • Search Google Scholar
    • Export Citation
  • Lau, K. M., K-M. Kim, and C. Hsu, 2005: Observational evidence of effects of absorbing aerosols on seasonal-to-interannual anomalies of the Asian monsoon. CLIVAR Exchanges, No. 3, International CLIVAR Project Office, Southampton, United Kingdom, 7–9.

    • Search Google Scholar
    • Export Citation
  • Li, J., W. P. Menzel, Z. Yang, R. A. Frey, and S. A. Ackerman, 2003: High-spatial-resolution surface and cloud type classification from MODIS multispectral band measurements. J. Appl. Meteor., 42 , 204226.

    • Search Google Scholar
    • Export Citation
  • Li, X., R. T. Pinker, M. M. Wonsick, and Y. Ma, 2007: Towards improved satellite estimates of short-wave radiative fluxes: Focus on cloud detection over snow. Part I: Methodology. J. Geophys. Res., 112 , D07208. doi:10.1029/2005JD006698.

    • Search Google Scholar
    • Export Citation
  • Li, Y., X. Liu, and B. Chen, 2006: Cloud type climatology over the Tibetan Plateau: A comparison of ISCCP and MODIS/TERRA measurements with surface observations. Geophys. Res. Lett., 33 , L17716. doi:10.1029/2006GL026890.

    • Search Google Scholar
    • Export Citation
  • Liu, H., and R. T. Pinker, 2008: Radiative fluxes from satellites: Focus on aerosols. J. Geophys. Res., 113 , D08208. doi:10.1029/2007JD008736.

    • Search Google Scholar
    • Export Citation
  • Liu, H., R. T. Pinker, and B. Holben, 2005: A global view of aerosols from merged transport models, satellite, and ground observations. J. Geophys. Res., 110 , D10S15. doi:10.1029/2004JD004695.

    • Search Google Scholar
    • Export Citation
  • Liu, H., R. T. Pinker, M. Chin, B. Holben, and L. Remer, 2008: Synthesis of information on aerosol optical properties. J. Geophys. Res., 113 , D07206. doi:10.1029/2007JD008735.

    • Search Google Scholar
    • Export Citation
  • McGarry, M. M., and R. J. Reed, 1978: Diurnal variations in convective activity and precipitation during Phases II and III of GATE. Mon. Wea. Rev., 106 , 101113.

    • Search Google Scholar
    • Export Citation
  • Murakami, M., 1983: Analysis of the deep convective activity over the western Pacific and Southeast Asia. Part I: Diurnal variation. J. Meteor. Soc. Japan, 61 , 6077.

    • Search Google Scholar
    • Export Citation
  • Nesbitt, S. W., and E. J. Zipser, 2003: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements. J. Climate, 16 , 14561475.

    • Search Google Scholar
    • Export Citation
  • NOAA/NESDIS/OSDPD/SSD, cited. 2008: IMS daily Northern Hemisphere snow and ice analysis at 4 km and 24 km resolution. National Snow and Ice Data Center, Boulder, CO, digital media. [Available online at http://nsidc.org/data/g02156.html].

    • Search Google Scholar
    • Export Citation
  • Parthasarathy, B., A. A. Munot, and D. R. Kothawale, 1995: Monthly and seasonal rainfall series for all-India homogeneous regions and meteorological subdivisions: 1871–1994. Research Rep. RR-065, Indian Institute of Tropical Meteorology, 113 pp.

    • Search Google Scholar
    • Export Citation
  • Patra, P. K., S. K. Behera, J. R. Herman, S. Maksyutov, H. Akimoto, and T. Yamagata, 2005: The Indian summer monsoon rainfall: Interplay of coupled dynamics, radiation, and cloud microphysics. Atmos. Chem. Phys., 5 , 21812188.

    • Search Google Scholar
    • Export Citation
  • Pinker, R. T., and Coauthors, 2003: Surface radiation budgets in support of the GEWEX Continental-Scale International Project (GCIP) and the GEWEX Americas Prediction Project (GAPP), including the North American Land Data Assimilation System (NLDAS) Project. J. Geophys. Res., 108 , 8844. doi:10.1029/2002JD003301.

    • Search Google Scholar
    • Export Citation
  • Pinker, R. T., X. Li, W. Meng, and E. Yegorova, 2007: Towards improved satellite estimates of short-wave radiative fluxes: Focus on cloud detection over snow. 2: Results. J. Geophys. Res., 112 , D09204. doi:10.1029/2005JD006699.

    • Search Google Scholar
    • Export Citation
  • Ramanathan, V., and Coauthors, 2001: Indian Ocean Experiment: An integrated analysis of the climate forcing and effects of the great Indo-Asian haze. J. Geophys. Res., 106 , (D22). 2837128398.

    • Search Google Scholar
    • Export Citation
  • Ramsay, B., 1998: The interactive multi-sensor snow and ice mapping system. Hydrol. Processes, 12 , 15371546.

  • Roca, R., and V. Ramanathan, 2000: Scale dependence of monsoonal convective systems over the Indian Ocean. J. Climate, 13 , 12861298.

  • Roca, R., S. Louvet, L. Picon, and M. Desbois, 2005: A study of convective systems, water vapor and top of the atmosphere cloud radiative forcing over the Indian Ocean using INSAT-1B and ERBE data. Meteor. Atmos. Phys., 90 , 4965.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and R. A. Schiffer, 1991: ISCCP cloud data products. Bull. Amer. Meteor. Soc., 72 , 220.

  • Rossow, W. B., and L. C. Garder, 1993a: Cloud detection using satellite measurements of infrared and visible radiances for ISCCP. J. Climate, 6 , 23412369.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and L. C. Garder, 1993b: Validation of ISCCP cloud detections. J. Climate, 6 , 23702393.

  • Rossow, W. B., and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80 , 22612287.

  • Rossow, W. B., A. W. Walker, and L. C. Garder, 1993: Comparison of ISCCP and other cloud amounts. J. Climate, 6 , 23942418.

  • Rossow, W. B., A. W. Walker, D. E. Beuschel, and M. D. Roiter, 1996: International Satellite Cloud Climatology Project (ISCCP) documentation of new cloud datasets. WMO/TD 737, World Meteorological Organization, 115 pp.

    • Search Google Scholar
    • Export Citation
  • Rozendaal, M. A., C. B. Leovy, and S. A. Klein, 1995: An observational study of diurnal variations of marine stratiform cloud. J. Climate, 8 , 17951809.

    • Search Google Scholar
    • Export Citation
  • Schiffer, R. A., and W. B. Rossow, 1983: The International Satellite Cloud Climatology Project (ISCCP): The first project of the World Climate Research Programme. Bull. Amer. Meteor. Soc., 64 , 779784.

    • Search Google Scholar
    • Export Citation
  • Schiffer, R. A., and W. B. Rossow, 1985: ISCCP global radiance data set: A new resource for climate research. Bull. Amer. Meteor. Soc., 66 , 14981505.

    • Search Google Scholar
    • Export Citation
  • Sen Roy, S., and R. C. Balling, 2007: Diurnal variations in summer season precipitation in India. Int. J. Climatol., 27 , 969976.

  • Shyamala, B., and C. V. V. Bhadram, 2006: Impact of mesoscale–synoptic scale interactions on the Mumbai historical rain event during 26–27 July 2005. Curr. Sci., 91 , 16491654.

    • Search Google Scholar
    • Export Citation
  • Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller, Eds. 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

    • Search Google Scholar
    • Export Citation
  • Sorooshian, S., X. Gao, K. Hsu, R. A. Maddox, Y. Hong, H. V. Gupta, and B. Imam, 2002: Diurnal variability of tropical rainfall retrieved from combined GOES and TRMM satellite information. J. Climate, 15 , 9831001.

    • Search Google Scholar
    • Export Citation
  • Sperber, K. R., and T. N. Palmer, 1996: Interannual tropical rainfall variability in general circulation model simulations associated with Atmospheric Model Intercomparison Project. J. Climate, 9 , 27272750.

    • Search Google Scholar
    • Export Citation
  • Stowe, L. L., P. A. Davis, and E. P. McClain, 1999: Scientific basis and initial evaluation of the CLAVR-1 Global Clear/Cloud Classification Algorithm for the Advance Very High Resolution Radiometer. J. Atmos. Oceanic Technol., 16 , 656681.

    • Search Google Scholar
    • Export Citation
  • Sui, C. H., K. M. Lau, Y. N. Takayabu, and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE. J. Atmos. Sci., 54 , 639655.

    • Search Google Scholar
    • Export Citation
  • Thomas, S. M., A. K. Heidinger, and M. J. Pavalonis, 2004: Comparison of NOAA’s operational AVHRR-derived cloud amount to other satellite-derived cloud climatologies. J. Climate, 17 , 48054822.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131 , 29613012.

  • Wielicki, B. A., and L. Parker, 1992: On the determination of cloud cover from satellite sensors: The effect of sensor spatial resolution. J. Geophys. Res., 97 (12) 1279912823.

    • Search Google Scholar
    • Export Citation
  • Yang, G. Y., and J. Slingo, 2001: The diurnal cycle in the tropics. Mon. Wea. Rev., 129 , 784801.

  • Yeh, T. C., and Y. X. Gao, 1979: The Meteorology of the Qinghai-Xizang (Tibet) Plateau. Science Press, 278 pp.

  • Zuidema, P., 2003: Convective clouds over the Bay of Bengal. Mon. Wea. Rev., 131 , 780798.

Fig. 1.
Fig. 1.

Meteosat-5 vs MODIS monthly mean total cloud amount (%) averaged zonally over longitudes 51°–95°E for (a) April and (b) July 2003.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 2.
Fig. 2.

Meteosat-5 vs ISCCP monthly mean total daytime cloud amount (%) averaged zonally over longitudes 51°–95°E for (a) April and (b) July 2003.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 3.
Fig. 3.

Average daytime cloud amount (%) for (top) premonsoon, (middle) peak-monsoon, and (bottom) postmonsoon seasons for (left) 2001 and (right) 2003.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 4.
Fig. 4.

Difference in monthly mean cloud amounts (%) for daylight hours for (a) June − May, (b) August − July, and (c) October − September 2001.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 5.
Fig. 5.

Amplitude of diurnal cycle of (left) low-cloud and (right) high-cloud amount averaged over the (a),(b) premonsoon, (c),(d) peak-monsoon, and (e),(f) postmonsoon seasons.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 6.
Fig. 6.

Location of data points analyzed in Figs. 7, 8, and 12.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 7.
Fig. 7.

Diurnal variation of total cloud amount (%) for (a) premonsoon, (b) peak-monsoon, and (c) postmonsoon season 2001 for the hours of 0800–1500 LST for points shown in Fig. 6.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 8.
Fig. 8.

Diurnal variation of frequency of occurrence of (left) low cloud and (right) high cloud for (top) premonsoon, (middle) peak-monsoon, and (bottom) postmonsoon seasons 2001 for the hours of 0800–1500 LST for points shown in Fig. 6.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 9.
Fig. 9.

Frequency of occurrence of deep convective clouds at selected daytime hours for the peak-monsoon season 2001.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 10.
Fig. 10.

Mean temperature of convection (K) for selected daylight hours for the peak-monsoon season 2001.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 11.
Fig. 11.

Total cloud amount (%) at 0600 UTC 1 Jul 2001 from (a) Meteosat-5 cloud analysis and (b) ERA-40; Meteosat-5 image from (c) IR channel and (d) visible channel.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Fig. 12.
Fig. 12.

Comparison of monthly mean total cloud amount (%) at 0600 UTC for each month of the 2001 monsoon season for Meteosat-5, ERA-40, NCEP–NCAR reanalysis, and ISCCP for each point in Fig. 6.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2027.1

Table 1.

Comparison of cloud tests used in cloud screening process for CCSDA vs Meteosat-5 processing.

Table 1.
Table 2.

Comparison of daytime average total cloud amount (%) from Meteosat-5 (upscaled to 2.5° resolution) and ISCCP D1 data at the six points shown in Fig. 6. Cloud amounts averaged for the 2001 pre-, peak-, and postmonsoon seasons are shown.

Table 2.
Save
  • Ackerman, S. A., R. E. Holz, R. Frey, E. W. Eloranta, B. C. Maddux, and M. McGill, 2008: Cloud detection with MODIS. Part II: Validation. J. Atmos. Oceanic Technol., 25 , 10731086.

    • Search Google Scholar
    • Export Citation
  • Babu, S. S., K. K. Moorthy, and S. K. Satheesh, 2004: Aerosol black carbon over Arabian Sea during intermonsoon and summer monsoon seasons. Geophys. Res. Lett., 31 , L06104. doi:10.1029/2003GL018716.

    • Search Google Scholar
    • Export Citation
  • Barros, A. P., G. Kim, E. Williams, and S. W. Nesbitt, 2004: Probing orographic controls in the Himalayas during the monsoon using satellite imagery. Nat. Hazards Earth Syst. Sci., 4 , 2951.

    • Search Google Scholar
    • Export Citation
  • Bergman, J. W., and M. L. Salby, 1996: Diurnal variations of cloud cover and their relationship to climatological conditions. J. Climate, 9 , 28022820.

    • Search Google Scholar
    • Export Citation
  • Brest, C. L., and W. B. Rossow, 1992: Radiometric calibration and monitoring of NOAA AVHRR data for ISCCP. Int. J. Remote Sens., 13 , 235273.

    • Search Google Scholar
    • Export Citation
  • Cherchi, A., and A. Navarra, 2003: Reproducibility and predictability of Asian summer monsoon in the ECHAM4-GCM. Climate Dyn., 20 , 365379.

    • Search Google Scholar
    • Export Citation
  • Chevallier, F., and G. Kelly, 2002: Model clouds as seen from space: Comparison with geostationary imagery in the 11-μm window channel. Mon. Wea. Rev., 130 , 712722.

    • Search Google Scholar
    • Export Citation
  • Chung, C., V. Ramanathan, and J. T. Kiehl, 2002: Effects of the South Asian absorbing haze on the northeast monsoon and surface–air heat exchange. J. Climate, 15 , 24622476.

    • Search Google Scholar
    • Export Citation
  • Dai, A., 2001: Global precipitation and thunderstorm frequencies. Part II: Diurnal variations. J. Climate, 14 , 11121128.

  • Frey, R. A., S. A. Ackerman, Y. Liu, K. I. Strabala, H. Zhang, J. R. Key, and X. Wang, 2008: Cloud detection with MODIS. Part I: Improvements in the MODIS cloud mask for collection 5. J. Atmos. Oceanic Technol., 25 , 10571072.

    • Search Google Scholar
    • Export Citation
  • Gadgil, S., 2003: The Indian Monsoon and its variability. Annu. Rev. Earth Planet. Sci., 31 , 429467.

  • Gambheer, A. V., and G. S. Bhat, 2001: Diurnal variation of deep cloud systems over the Indian region using INSAT-1B pixel data. Meteor. Atmos. Phys., 78 , 215226.

    • Search Google Scholar
    • Export Citation
  • Govaerts, Y. M., M. Clerici, and N. Clerbaux, 2004: Operational calibration of the Meteosat radiometer VIS band. IEEE Trans. Geosci. Remote Sens., 42 , 19001914.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., and R. W. Jacobson, 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105 , 11711188.

  • Hou, Y. T., K. A. Campana, K. E. Mitchell, S. K. Yang, and L. L. Stowe, 1993: Comparison of an experimental NOAA AVHRR cloud dataset with other observed and forecast cloud datasets. J. Atmos. Oceanic Technol., 10 , 833849.

    • Search Google Scholar
    • Export Citation
  • Islam, M. N., T. Hayashi, H. Uyeda, T. Terao, and K. Kikuchi, 2004: Diurnal variations of cloud activity in Bangladesh and north of the Bay of Bengal in 2000. Remote Sens. Environ., 90 , 378388.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., P. A. Arkin, and M. Morrissey, 1994: An examination of the diurnal cycle in oceanic tropical rainfall using satellite and in situ data. Mon. Wea. Rev., 122 , 22962311.

    • Search Google Scholar
    • Export Citation
  • Kistler, R., and Coauthors, 2001: The NCEP–NCAR 50-Rear Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247267.

    • Search Google Scholar
    • Export Citation
  • Krishnamurthy, V., and J. Shukla, 2000: Intraseasonal and interannual variability of rainfall over India. J. Climate, 13 , 43664377.

  • Krishnamurti, T. N., and C. M. Kishtawal, 2000: A pronounced continental-scale diurnal mode of the Asian summer monsoon. Mon. Wea. Rev., 128 , 462473.

    • Search Google Scholar
    • Export Citation
  • Lau, K. M., K-M. Kim, and C. Hsu, 2005: Observational evidence of effects of absorbing aerosols on seasonal-to-interannual anomalies of the Asian monsoon. CLIVAR Exchanges, No. 3, International CLIVAR Project Office, Southampton, United Kingdom, 7–9.

    • Search Google Scholar
    • Export Citation
  • Li, J., W. P. Menzel, Z. Yang, R. A. Frey, and S. A. Ackerman, 2003: High-spatial-resolution surface and cloud type classification from MODIS multispectral band measurements. J. Appl. Meteor., 42 , 204226.

    • Search Google Scholar
    • Export Citation
  • Li, X., R. T. Pinker, M. M. Wonsick, and Y. Ma, 2007: Towards improved satellite estimates of short-wave radiative fluxes: Focus on cloud detection over snow. Part I: Methodology. J. Geophys. Res., 112 , D07208. doi:10.1029/2005JD006698.

    • Search Google Scholar
    • Export Citation
  • Li, Y., X. Liu, and B. Chen, 2006: Cloud type climatology over the Tibetan Plateau: A comparison of ISCCP and MODIS/TERRA measurements with surface observations. Geophys. Res. Lett., 33 , L17716. doi:10.1029/2006GL026890.

    • Search Google Scholar
    • Export Citation
  • Liu, H., and R. T. Pinker, 2008: Radiative fluxes from satellites: Focus on aerosols. J. Geophys. Res., 113 , D08208. doi:10.1029/2007JD008736.

    • Search Google Scholar
    • Export Citation
  • Liu, H., R. T. Pinker, and B. Holben, 2005: A global view of aerosols from merged transport models, satellite, and ground observations. J. Geophys. Res., 110 , D10S15. doi:10.1029/2004JD004695.

    • Search Google Scholar
    • Export Citation
  • Liu, H., R. T. Pinker, M. Chin, B. Holben, and L. Remer, 2008: Synthesis of information on aerosol optical properties. J. Geophys. Res., 113 , D07206. doi:10.1029/2007JD008735.

    • Search Google Scholar
    • Export Citation
  • McGarry, M. M., and R. J. Reed, 1978: Diurnal variations in convective activity and precipitation during Phases II and III of GATE. Mon. Wea. Rev., 106 , 101113.

    • Search Google Scholar
    • Export Citation
  • Murakami, M., 1983: Analysis of the deep convective activity over the western Pacific and Southeast Asia. Part I: Diurnal variation. J. Meteor. Soc. Japan, 61 , 6077.

    • Search Google Scholar
    • Export Citation
  • Nesbitt, S. W., and E. J. Zipser, 2003: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements. J. Climate, 16 , 14561475.

    • Search Google Scholar
    • Export Citation
  • NOAA/NESDIS/OSDPD/SSD, cited. 2008: IMS daily Northern Hemisphere snow and ice analysis at 4 km and 24 km resolution. National Snow and Ice Data Center, Boulder, CO, digital media. [Available online at http://nsidc.org/data/g02156.html].

    • Search Google Scholar
    • Export Citation
  • Parthasarathy, B., A. A. Munot, and D. R. Kothawale, 1995: Monthly and seasonal rainfall series for all-India homogeneous regions and meteorological subdivisions: 1871–1994. Research Rep. RR-065, Indian Institute of Tropical Meteorology, 113 pp.

    • Search Google Scholar
    • Export Citation
  • Patra, P. K., S. K. Behera, J. R. Herman, S. Maksyutov, H. Akimoto, and T. Yamagata, 2005: The Indian summer monsoon rainfall: Interplay of coupled dynamics, radiation, and cloud microphysics. Atmos. Chem. Phys., 5 , 21812188.

    • Search Google Scholar
    • Export Citation
  • Pinker, R. T., and Coauthors, 2003: Surface radiation budgets in support of the GEWEX Continental-Scale International Project (GCIP) and the GEWEX Americas Prediction Project (GAPP), including the North American Land Data Assimilation System (NLDAS) Project. J. Geophys. Res., 108 , 8844. doi:10.1029/2002JD003301.

    • Search Google Scholar
    • Export Citation
  • Pinker, R. T., X. Li, W. Meng, and E. Yegorova, 2007: Towards improved satellite estimates of short-wave radiative fluxes: Focus on cloud detection over snow. 2: Results. J. Geophys. Res., 112 , D09204. doi:10.1029/2005JD006699.

    • Search Google Scholar
    • Export Citation
  • Ramanathan, V., and Coauthors, 2001: Indian Ocean Experiment: An integrated analysis of the climate forcing and effects of the great Indo-Asian haze. J. Geophys. Res., 106 , (D22). 2837128398.

    • Search Google Scholar
    • Export Citation
  • Ramsay, B., 1998: The interactive multi-sensor snow and ice mapping system. Hydrol. Processes, 12 , 15371546.

  • Roca, R., and V. Ramanathan, 2000: Scale dependence of monsoonal convective systems over the Indian Ocean. J. Climate, 13 , 12861298.

  • Roca, R., S. Louvet, L. Picon, and M. Desbois, 2005: A study of convective systems, water vapor and top of the atmosphere cloud radiative forcing over the Indian Ocean using INSAT-1B and ERBE data. Meteor. Atmos. Phys., 90 , 4965.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and R. A. Schiffer, 1991: ISCCP cloud data products. Bull. Amer. Meteor. Soc., 72 , 220.

  • Rossow, W. B., and L. C. Garder, 1993a: Cloud detection using satellite measurements of infrared and visible radiances for ISCCP. J. Climate, 6 , 23412369.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and L. C. Garder, 1993b: Validation of ISCCP cloud detections. J. Climate, 6 , 23702393.

  • Rossow, W. B., and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80 , 22612287.

  • Rossow, W. B., A. W. Walker, and L. C. Garder, 1993: Comparison of ISCCP and other cloud amounts. J. Climate, 6 , 23942418.

  • Rossow, W. B., A. W. Walker, D. E. Beuschel, and M. D. Roiter, 1996: International Satellite Cloud Climatology Project (ISCCP) documentation of new cloud datasets. WMO/TD 737, World Meteorological Organization, 115 pp.

    • Search Google Scholar
    • Export Citation
  • Rozendaal, M. A., C. B. Leovy, and S. A. Klein, 1995: An observational study of diurnal variations of marine stratiform cloud. J. Climate, 8 , 17951809.

    • Search Google Scholar
    • Export Citation
  • Schiffer, R. A., and W. B. Rossow, 1983: The International Satellite Cloud Climatology Project (ISCCP): The first project of the World Climate Research Programme. Bull. Amer. Meteor. Soc., 64 , 779784.

    • Search Google Scholar
    • Export Citation
  • Schiffer, R. A., and W. B. Rossow, 1985: ISCCP global radiance data set: A new resource for climate research. Bull. Amer. Meteor. Soc., 66 , 14981505.

    • Search Google Scholar
    • Export Citation
  • Sen Roy, S., and R. C. Balling, 2007: Diurnal variations in summer season precipitation in India. Int. J. Climatol., 27 , 969976.

  • Shyamala, B., and C. V. V. Bhadram, 2006: Impact of mesoscale–synoptic scale interactions on the Mumbai historical rain event during 26–27 July 2005. Curr. Sci., 91 , 16491654.

    • Search Google Scholar
    • Export Citation
  • Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller, Eds. 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

    • Search Google Scholar
    • Export Citation
  • Sorooshian, S., X. Gao, K. Hsu, R. A. Maddox, Y. Hong, H. V. Gupta, and B. Imam, 2002: Diurnal variability of tropical rainfall retrieved from combined GOES and TRMM satellite information. J. Climate, 15 , 9831001.

    • Search Google Scholar
    • Export Citation
  • Sperber, K. R., and T. N. Palmer, 1996: Interannual tropical rainfall variability in general circulation model simulations associated with Atmospheric Model Intercomparison Project. J. Climate, 9 , 27272750.

    • Search Google Scholar
    • Export Citation
  • Stowe, L. L., P. A. Davis, and E. P. McClain, 1999: Scientific basis and initial evaluation of the CLAVR-1 Global Clear/Cloud Classification Algorithm for the Advance Very High Resolution Radiometer. J. Atmos. Oceanic Technol., 16 , 656681.

    • Search Google Scholar
    • Export Citation
  • Sui, C. H., K. M. Lau, Y. N. Takayabu, and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE. J. Atmos. Sci., 54 , 639655.

    • Search Google Scholar
    • Export Citation
  • Thomas, S. M., A. K. Heidinger, and M. J. Pavalonis, 2004: Comparison of NOAA’s operational AVHRR-derived cloud amount to other satellite-derived cloud climatologies. J. Climate, 17 , 48054822.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131 , 29613012.

  • Wielicki, B. A., and L. Parker, 1992: On the determination of cloud cover from satellite sensors: The effect of sensor spatial resolution. J. Geophys. Res., 97 (12) 1279912823.

    • Search Google Scholar
    • Export Citation
  • Yang