CHAPTER XIV.

FORECASTING FOR SOLITARY OBSERVERS.

NATURE OF THE PROBLEM.

A COMPREHENSIVE view of weather-science divides itself into three problems-one direct and two inverse. The direct problem of weather is to explain by mechanical causes the origin and nature of all the complicated phenomena of wind and weather which present themselves to our senses, and the nature of the sequence of weatherchanges. This we have already partially done in the preceding pages. The inverse problem of meteorology is, given a portion of a sequence of the weather, to tell what is going to follow. The morning is fine, but now cirrus begins to form, and the mercury has begun to fall-what weather is coming? Last night a cyclone lay over Ireland, this morning it covers Wales-what will the weather be over Great Britain for the rest of the day?

These two illustrations point at once to a natural subdivision of the questions of forecasting: the best that a single observer can do, who has his eyes to look at the appearance of the sky and any instruments at his

disposal; and the best that a meteorologist can do, who is seated in a central bureau, with abundant telegraphic intelligence for many miles round the country for which he has to issue forecasts, so as to enable him to construct synoptic charts at such intervals as he may think necessary. The latter doubtless represents the highest development of which forecasting is capable; but the former can never be superseded for use among sailors, fishermen, and shepherds. For this reason we will discuss them in separate chapters, and we will take the problem of a solitary observer first, as it is the older and the more generally useful. We shall only attempt to give general principles, and not to go into all the details for any one country.

PROGNOSTICS.

We have already gone very fully into the subject of prognostics, and pointed out both the reasons for their success as well as for their failure. When we come to look at all that has been done, we see that, on the whole, we have not been able to develop the practical utility of prognostics very materially, though we have been able to place the whole branch of the subject on a scientific basis.

The most valuable addition of recent times to weatherlore is undoubtedly in the methodical observation of cirrus clouds. The recognition of cirrus as a sign of rain is as old as meteorology, but the deductions which can be made from the direction of the motion of the upper clouds are quite of modern date. No absolute test can

be given for the discrimination of fine weather from dangerous cirrus beyond the general surroundings and experience of the observer; but Ley has shown the importance of noting by eye the velocity of the cirrus, because rapid-moving cirrus is a much worse sign of the weather than slowly moving cloud. This is probably one of the most important advances which has been made.

THE BAROMETER.

We propose rather in this chapter to deal with the value of the indications which the barometer can afford to a solitary observer, and especially to explain why the indications of that instrument so often fail.

Why do we sometimes have rain with a rising or steady barometer, and why is the weather sometimes fine with a falling barometer? Then, again, why do we sometimes experience a heavy gale with only a slight fall of the mercury, while at other times the barometer will fall very low without any unusual amount of wind?

These apparent anomalies in the indications of the barometer occur all over the world, and those in each country must be explained by reference to the meteorology of the place. Though we shall draw our illustrations from Great Britain only, the principles which we shall lay down are of universal application. In no branch of the subject shall we find synoptic charts more indispensable, for without them no explanation could ever have been afforded of irregular barometric fluctuations.

GENERAL INDICATIONS.

The preceding chapters will have sufficiently explained the reasons for what we may call the generally correct indications of the barometer. We can now readily understand why the rapid rise in rear of a cyclone indicates unsettled weather, and the gradual rise of an incipient anticyclone settled fine weather; also why the steady barometer of a persistent anticyclone indicates dry seasonable weather, and the rapid fall of an oncoming cyclone presages storm and rain. All these indications of the barometer can be detected by intermittent observations, or, in fact, by merely looking occasionally at the instrument.

The author has, however, discovered that we can sometimes utilize the greater refinements of self-registered barographs to deduce some knowledge of the future force of the wind from flexures in the recorded curves. These deductions are of the more value now that efficient barographs are so cheap as to be within the reach of everybody.

AUTHOR'S RULES FOR INFERRING FROM A BAROGRAM WHETHER A GALE IS GOING TO INCREASE DECREASE.

OR

The principle on which the author's rules are founded depend on what is called the "direction of curvature" of a curve. In the lower portion of Fig. 84, the portion of the trace near the letter A has its hollow turned upwards, and is called convex, relative to the base line. A little

further down, near the figures 14, the curve is hollowed downwards, and would be called concave.

The other half
From this we

of the curve is convex almost throughout. see that both convexity and concavity are independent of

FIG. 84.-Gradients, and flexure of barogram.

whether the mercury is rising or falling, and also of the rapidity of their rise or fall.

If the barometer change at a uniform rate, either upwards or downwards, it is evident that the resulting trace will be a straight line, either rising or falling, and it does not the least matter how rapid the rise or fall is. If, however, the rate of fall changes with diminishing pressure, then the curve will become convex or concave,