15 Cloud Analysis and Forecasting

Sunayan Saha and Pragati Pramanik Maity

1. Learning outcomes
2. Introduction
3. Definitions: cloud cover, ceiling and visibility
4. Techniques for cloud observation and analysis
4.1. Visual observations
4.2. Interpretation of visible imageries from satellite
4.3. Interpretation of satellite IR imagery
5. Distribution of cloudiness
6. Forecasting of cloud
7. Summary

1. Learning outcomes

• After studying this module, you shall be able to:
• know the definitions of important terminologies and standard symbols related to clouds
• know the techniques to distinguish clouds from other features and differentiate among various cloud forms using satellite imageries
• know the distribution of cloudiness, zones of maximum and minimum cloudiness and precipitation across the globe
• know the steps involved in cloud forecasting

2. Introduction

Clouds are important within the climatic system as they interact with both short wave solar radiation and long wave terrestrial radiation. With recent estimates of the future increase in atmospheric CO2 concentration, climate models have predicted a rise in global mean temperature by about 2- 5 degree Kelvin by the turn of the next century (IPCC, 2007). It is well known that much of the uncertainty in the global mean temperature rise is due to the frequency and behaviour of clouds in a future climate (IPCC, 2007). Apart from its importance in the climatic system, clouds are also important in short-term forecasts produced by numerical weather prediction models as good forecasts of cloud position and type are the first step towards reliable forecasts on precipitation location and intensity. Another reason of importance of cloud in weather forecasts is that low level clouds tend to reduce visibility of mountains and tall obstructions like trees & buildings. This can affect the flight patterns of civilian and military aircrafts as well as military and rescue operations. Clouds influence the radiation balance of earth and hence its correct representation is important to ensure accurate surface temperature forecasts; in warm conditions this will affect the rate of evaporation and soil water status, while under cold conditions the amount of cloud is important for frost forecasting. Lastly, the amount of ice formation on aircrafts while they cruise through clouds is dependent on the water content, phase and temperature of the clouds, so accurate forecast of clouds is important to ensure safe flight conditions. Following sections deal with various techniques of cloud analysis and foresting and explain important terminologies associated with them. The module also delves upon the cloudiness and precipitation pattern across the globe which has emerged as a result of long-term meteorological observations and coordinated research.

3. Definitions: cloud cover, ceiling and visibility

Clouds can be differentiated from other earth surface features based on its higher reflectance

values and lower temperature. Hence, the threshold approaches based on the visible and infrared bands can successfully detect the cloud in normal surface conditions. However, surface conditions where snow and/or ice are present along with clouds, the simple threshold approach may not be effective. In addition, it is challenging to detect some of the cloud types, viz., cirrus, low stratus and cumulus due to their insufficient contrast with the surface. Moreover, the partly cloudy and partly clear views in the satellite images make the demarcation of the cloud edges more difficult. Of late, with the advancement in the satellite and camera technology, the challenges in the cloud detection can be overcome using the multispectral approaches.

The cloud cover is estimated by the observer and expressed on a scale ranging from 0 to 8. Here, 0 means clear sky, 4 means that one half (or four-eighths) of the sky is covered, and 8 means overcast. Till 1948, the cloud cover was used to be reported on a scale ranging from 0 to 10, and hence while analyzing the past records it is necessary to ascertain which scale was in use.

The ceiling or the height above the ground of the cloud cover is of particular importance in aircraft operation. A ceiling is said to exist if the clouds below 10,000 ft cover more than one half (4/8) of the sky. If the cloud cover is less, or the base higher, the ceiling is said to be unlimited. The height of the ceiling is measured by small balloons whose ascensional velocity is known. At night, it may be measured by attaching a light to the balloon. More commonly used is the “ceiling light”, which is a vertical beam that throws a spot on the base of the cloud. The elevation angle of the spot is observed from a neighbouring place and the height is obtained by simple trigonometry.

The visibility is the largest distance at which prominent objects, such as mountains hills, buildings, towers etc can be identified by the unaided eye. The distance depends upon the impurities in the air and upon the density of fog, mist, precipitation, and so forth. In a dense smog the visibility may be only a few feet, while in the pure air of the arctic winter, mountains may be seen at a distance of more than 100 miles.

4. Techniques for cloud observation and analysis

• 1. Visual observations

Elements such as cloud cover, cloud forms and forms of precipitations etc are indicated by symbols which are intended to remind the reader of the phenomenon to which they refer. Cloud cover is estimated by the observer and reported as the number of eighths of the sky covered by clouds as shown in table 1. The station is identified by a circle and the cloud cover is indicated by proportional filling:

The unit of cloud cover at any given location such as a weather station is termed as an okta. Zero (0) okta means completely clear sky whereas 8 oktas indicate a completely overcast situation. In addition, in the SYNOP code there is an extra cloud cover indicator ‘9’ (as can be seen in table 1) meaning that the sky is totally obscured i.e. hidden from view, usually due to dense fog or heavy snow and hence the cloud cover cannot be determined.

After the cloud cover is quantified and reported, next we need to know what type of clouds is there in the sky. The standard symbols of the different cloud types as per classification of clouds are contained in table 2.

Table 2: The primary symbols for principal cloud forms

The above symbols may appear in combinations also such as cumulonimbus with scud, altocumulus with altostratus, cumulus and stratocumulus etc.

4.2. Interpretation of visible imageries from satellite

Satellite imagery is being extensively used by synoptic network in conjunction with other available conventional meteorological data for analysis and weather forecasting. Zones of cloudiness are identified from the satellite imagery as regions of upward velocity and hence potential areas for occurrence of rainfall. Visible, infra red and water vapour images have distinctive uses and are complementary to each other. The satellite cloud images, recorded by infrared and visible channel, mainly reflect the top brightness temperature and albedo information of clouds. Due to its higher albedo than other earth surface features except snow cover, in visible band imagery the clouds appear as white or light grey. Based on the variation of albedo in visible imagery, the cloud, land and sea surface can be easily differentiated. The sea and other water bodies have lower albedo and look darker, whereas land appears relatively brighter than sea but darker than clouds due to its intermittent albedo values.

The physical properties of the clouds, i.e. depth, water content, cloud-droplet size etc. affect the albedo which ultimately reflects in the colour and brightness of the clouds. In case of larger depth, high cloud water and small cloud-droplet size, the cloud albedo values will be high and vice-versa. Moreover, as the sun shines obliquely on the cloud structure, the shadow of the upper part of the cloud falls on the lower part and as a result the vertical structure of the cloud is revealed in visible band imageries. Along with the detection, the identification of different types of clouds can be possible by utilizing the textural information, e.g. stratocumulus clouds (Sc) can be distinguished from stratus (St) by using the cellular pattern in visible imageries. The typical albedo values for different earth surface features and clouds are given in table 3.

Though visible imagery is easy to interpret, it has some limitations, like visible imagery cannot be obtained during night hours; to distinguish clouds from snow cover, knowledge of the surface topography is essential; the appearance of mesoscale clouds may be different from actual when the cloud size is smaller than spatial resolution of the imagery. Moreover, based on the background surface the thin clouds may be over or underestimated, e.g. due to low albedo of thin clouds it may not be properly detected over dark surfaces, whereas over bright surface like desert it may appear misleadingly brighter and thicker.

Table 3: Typical albedo values (%) for earth surface features and clouds

3.4.        Interpretation of satellite IR imagery

In infrared imageries, due to their lower temperature the clouds generally appear lighter than other earth surface features. In this regard, both the infrared and visible imageries have the similarity.

As the cloud temperature is decreasing with increasing height, the contrast among different levels of cloud can be visualized in infrared imageries, but not in visible imageries. Similarly, the thin transparent cirrus cloud over warmer surface can be detected in infrared images, whereas it is not possible with visible band alone. Along with all these advantages of infrared imagery over visible, it also has some limitations. Unlike visible, it is not possible to extract cloud textural information from infrared imageries as infrared is based on emitted radiation not scatted radiation. Moreover, during night hours the low clouds and fog cannot be detected as the temperature difference with the underlying surface is very less. Satellite imageries derived from different bands of INSAT-3D are shown in Fig.1.

In the above figure, warmest (lowest) clouds are shown in white whereas coldest (highest) clouds are displayed in shades of yellow, red and purple. Geostationary satellites like INSAT, GOES or METEOSAT- series satellites are normally deployed for real time monitoring of cloud and some other weather parameters.

Given below are some general rules to determine cloud characteristics when comparing visible and infrared satellite images:

1. If the cloud is bright white on infrared then it is a high cloud or has a cloud top that is developed high into the troposphere.
2. If a cloud is bright white on visible but is not bright enough or hard to see on infrared image then it is likely to be a cloud occurring close to the earth’s surface. This can happen when there is a thick layer of fog or stratus cloud near the surface.
3. Thunderstorms appear bright white on both visible and infrared bands. A thick cloud appears bright white in visible band due to high albedo and cold cloud tops are assigned bright white colour on infrared images. To be sure about the presence of a thunderstorm in an area some other features should be looked at like anvil blow-off, overshooting top and extremely textured appearance in visible imagery.
4. If a cloud is not very white on visible as well as IR imageries then it is likely a thin cloud or a cloud occurring near the surface.
5. Around the sunset time there is less reflection of sun rays from the cloud, hence in visible images clouds will not appear as white as it would be in other times of the day with different incidence angles.
6. High level clouds (e.g. cirrus) or detached anvils of thunder clouds look wispy in visible channel and very bright white in infrared channel.
7. Cumulus clouds have a lumpy texture. Stratus clouds have a flat texture especially on infrared. Cirrus clouds tend to be thin and show up white on infrared.

In Fig. 2 below, a comparison has been made between a VIS and IR image of the same region.

In the above figure, it can be seen that much of the cloud is bright white on VIS in Texas but have much darker shade on IR imagery. This indicates low clouds. There are thunderstorms in eastern Tennessee as can be seen that clouds there are bright white in both VIS and IR images. Around the western Mississippi cloud areas are less brighter than the thunderstorm clouds but are more brighter or whiter than the low lying clouds in much of Texas on IR image. Hence, it can be inferred that the clouds of western Mississippi are middle to high level clouds and could be producing light precipitation.

4. Distribution of cloudiness

There are two zones of maximum cloudiness in each hemisphere. The primary maxima of cloudiness are found in the temperate zones, between 30 – 60o N and S latitudes where there is prevalence of cyclones and fronts. The secondary maxima of cloudiness are positioned in the equatorial region. However, in terms of rainfall, the former shows the minima and the later the maxima. The reason for such contrast in cloudiness and rainfall behavior is that in temperate zone, precipitation occurs mainly from the stratus type of clouds, which although covers the whole or most part of the sky, produces only light or moderate rains. Whereas, in the equatorial belt most clouds attain great vertical developments (e.g. cumulus, cumulonimbus), although localized but they produce heavy intensity rains.

In each hemisphere, there is one belt of the minimum amount of cloudiness. This coincides with the latitudes of subtropical anticyclones and trade winds. Within the subtropical high pressure belt the cloudiness is less on the continents than on the oceans. Because of the descending air currents in the subtropical belt, all the hot deserts of the world are located here. It is, therefore, natural for cloudiness to reach its minimum in these regions marked by clear skies.

Besides, there are annual as well as diurnal variations in the amount of cloudiness in a particular region. In the equatorial region, there is little variation in the amount of cloudiness from one month to the other. But between 10 degree and 20 degree North and South latitudes the cloud maximum occurs during summer months. This is the period of maximum rainfall also. The western sides of continents within the subtropical belt have the cool season maxima of cloudiness as well as precipitation. In the higher latitudes, the continental interiors have summer maxima of cloudiness, since winters are marked by anticyclonic conditions there.

According to Trewartha, daily variation in cloudiness depends on the type of cloud present in the sky. In case of cumulus or cumuliform clouds, a maximum is reached in the early and middle afternoon. Stratus and other stratiform clouds show their maximum in the early morning.

5. Forecasting of cloud

Depending on the time period of validity, cloud forecasting can be grouped into short term- and long term. In short term, validity period is maximum of about 12 hours and in long term it is beyond 12 hours. In case of frontal clouds, for both type of forecasting, three major aspects need careful considerations. These are given below:

Short-term forecasting:

1. current cloud situation analysis: With the help of combinations of surface observations, upper air observations and ascents and satellite imageries current cloud situation can be analyzed in detail. Only surface observations may not be sufficient as higher level clouds may be blocked from view/detection. On the other hand, satellites sometimes can’t detect the lower clouds due to the shadow of upper level clouds on the lower level clouds.
2. Advection of the current state of cloud with wind field: This involves analysis of the component of gradient wind normal to front and the actual or forecast wind situation at cloud level
3. Development of the current state with time: To know this we need to find out the pressure tendencies, any wave development on front, development areas on upper air charts, local effects such as orographic barrier, nearness to the coast etc. Early frames of model output are verified against actual observations for refining the forecasts.

Long- term forecasting:

The guidelines for short-term forecasting are also applicable for this type of forecasting. However, forecast reliability becomes less and less with time. In addition to the above, model outputs are used to position fronts/ lows/rain-bands and assess the likely activity of the system. classical frontal theory (keeping in mind the limitation of the theory) and local knowledge are applied to produce a best estimate of the probable cloud structure.

6. Summary

This module can be summarized as follows:

• Clouds are important in the climatic system as they influence the radiation budget of any location on the earth.
• Good forecasts of cloud position and cloud type are the first step to achieve good forecasts on rainfall location and its intensity.
• Conventionally, the cloud cover is estimated from ground on a numerical scale of 0 to 8. Modern cloud detection techniques are based on the satellite imageries in the visible and infrared bands. Visible and infrared images have distinctive uses and are complementary to each other.
• Clouds are distinguished from other objects based on its higher reflectance values and lower temperature. Due to its higher albedo than other earth surface features except snow cover, clouds appear as white or light grey in a visible imagery.
• Infrared and visible channel images of cloud recorded by a satellite mainly reflect the top brightness temperature and albedo of clouds.
• The forecasting of frontal cloud involves three principal considerations, viz., analysis of the current cloud situation, advection of the current state of cloud with wind field and development of current state with time.
• Satellite imageries are being extensively used by synoptic network in conjunction with conventional meteorological data for cloud analysis and rainfall / weather forecasting.
• Daily variation in cloudiness depends on the type of cloud. For example, amount of cumulus or cumuliform clouds reach their maximum in the early and middle afternoon whereas for stratus and other stratiform clouds it is in the early morning.
• There are two zones of maximum cloudiness in each hemisphere. The primary maxima lie between 30o and 60o North and South latitudes. Equatorial region show the secondary maxima. However, in terms of rainfall, the former shows the minima and the later the maxima.
• Subtropical highs (STH) are the belts of minimum cloudiness in each hemisphere. Within the subtropical highs, cloudiness is less on the continents than on the oceans. The western sides of continents within STH belt have maximum cloudiness and precipitation during the cold season. In the higher latitudinal regions, the continental interiors have maximum cloudiness during summer.
• Near equator, there is least month to month variation in cloudiness. On the other hand, between 10 and 20 degree latitudes in both hemispheres the maximum cloud as well as rainfall occurs during summer months.