There is a popular Christmas carol that begins with these lines: Said the night wind to the little lamb, ‘Do you see what I see?’ The carol is about the forces of nature communicating with each other and rejoicing over the birth of the saviour. But the question can also be asked by one human being to another and that too in today’s world. A person with impaired vision sees the same thing differently from a person with a healthy eye. Even with two persons with normal vision, one may be fascinated by a particular object but another may fail to notice it. If our knowledge of the world depends upon how our senses react to it, then how can we be sure that what we observe is the truth or that it really exists? If we try to acquire knowledge with the help of instruments so that subjectivity is eliminated, further questions arise.

In most branches of science, the researchers conduct their own experiments and measurements are made by the scientists themselves or by team members working in their laboratory. Generally speaking, the experiments are repeatable and verifiable by other scientists if they want to. However, in the case of geophysical measurements, this is not possible. Meteorologists require global data for operational analysis and forecasting, for improving their understanding of the atmosphere and for carrying out climate research. For these purposes they are compelled to rely on measurements made in different countries, by people unknown to them, following different practices and conventions, and using different types of instruments and observational techniques. Geophysical measurements once made cannot be repeated or verified independently later on.

The entire science and profession of meteorology is built upon the assumption that meteorological observations reported from the world over are made properly, carefully and honestly. However, the additional question that remains is whether, or to what extent, they represent the truth about the state of the atmosphere at a given place and instant. Weather prediction involves the application of physical laws to the initial state of the atmosphere in order to derive its state later in time. So the accuracy of weather prediction depends upon how truthfully the observations represent the initial state.

Pontius Pilate, while interrogating a reticent Jesus before sentencing him to death on the cross, had asked “What is truth?” He had posed the question more to himself than to Jesus, and there is no record of his having answered it. Great philosophers have been struggling to find an answer to this very fundamental question, but we do not have it yet. “There is no truth. There is only perception”, said Gustave Flaubert, the great French novelist. But the problem is, what happens when I see but do not perceive, or when you do not see what I see?

Science is interested in facts and controlled observations. Such empiricism has its problems in meteorology where the observations cannot be made in controlled environments. The atmosphere cannot be measured at all times and places continuously but only at convenient times and places. This non-uniform sampling may or may not help in constructing a picture that is representative of the real atmosphere.

The instruments that we use for meteorological observations give us an idea about only those properties of the atmosphere that the instruments are capable of measuring. We need to employ several types of instruments and observational techniques, including remote sensing, that will together provide us a total picture of the atmosphere, but again how far that will represent the real atmosphere remains a question.

Moreover, in the case of meteorological observations, a fundamental problem is that the instruments that are employed for observation themselves affect or modify the properties of the atmosphere that they are supposed to measure. The process of meteorological observation by its very nature causes a distortion of truth, as the atmosphere would have been different had the instrument not been there at all. Meteorological instruments have therefore to be designed and deployed in such a way that there is a minimal disturbance to the natural atmosphere.

Let us take the example of the simple air thermometer. It has first to be compared against a standard thermometer and calibrated, but once that is done it can be used for measuring air temperature accurately as error corrections can be applied. But where in the atmosphere, optimally, do you keep an air thermometer? Not on open ground, as it will record the ground temperature. Not in the open air, as direct sunlight will heat it and the reading will be far higher than that of the air temperature, or it may get cooled by rain or snow. Not inside a closed room, as it will measure the temperature of the stagnant air within.

In the early days of IMD’s history, air temperature used to be recorded in a small hut which had a thatched roof to keep off sun and rain and was open from all sides to allow air currents to flow freely around the thermometer. Again, such a hut could be designed in many ways, and so IMD had a Bengal hut and a Madras hut!

Different countries had different practices of measuring air temperature until the 1920s when what came to be known as the Stevenson Screen was adopted internationally for placing thermometers. A Stevenson screen is essentially a box mounted on pillars at a height convenient to observers, which blocks sunlight from falling on the thermometers and has slits or louvers for the air to pass around them. However, while the Stevenson Screen became popular around the globe, no common design was universally accepted. The Stevenson Screens came with a great variety of height, size, internal layout, external shape, material used for construction, number of pillars and so on. Each of these factors influenced the reading of the thermometer within the screen and the result was that the temperature readings in differently designed screens were not strictly comparable.

Systematic international intercomparisons of thermometers have been regularly organised under the aegis of the World Meteorological Organization to ensure compatibility of temperature measurements made throughout the global observing network. However, the same is not the case as regards thermometer screens. The lack of standardization of Stevenson Screens and non-uniformity of their design continues to be a problem even today. In fact the problem has been compounded in recent years because of the automatic weather stations that are now coming into use in large numbers. The AWS sensors are small sized and so require smaller screens. The screens are made of light plastic material and can be installed at any height as an observer is not required to reach it himself. The specifications of AWS screens vary significantly from one manufacturer to another.

Only a few international intercomparisons of thermometer screens have been organized so far. Some of them have revealed that the temperatures recorded with identical thermometers installed in different types of screens could vary up to 0.2°C, but with certain types of screens under certain conditions, the air temperature differences could be large as 1°C or even more. This is a result that needs to be carefully considered while processing temperature data to compute the precise magnitude of global warming.

The observed rise in the earth’s average temperature over the previous 100 years is of the order of 0.7 °C. While analysing the past temperature data, it is necessary to isolate the errors that could be attributable to the dissimilar designs of thermometer screens in which the temperatures were recorded in the past.