Environmental forecasting

Introduction

The modern approach to environment forecasting really began when national meteorological services were established in Europe and the USA in the 1850's. These services were based on the improved empirical and scientific knowledge of meteorology, the coordination of existing networks of meteorological observers and the setting up of new communications links using the recently invented electric telegraph. Of course, organizations for forecasting natural phenomena had been in existence since the earliest times, notably the Chinese system for flood forecasting and warning.

But they did not have all the elements of a modern system. Amongst the meteorological pioneers were the astronomer Le Verrier in Paris, the scientist Buys-Ballot in the Netherlands and the hydrographers Maury in USA, and Fitzroy in London. Some of their innovations are still in use, such as Buys-Ballot's law and Fitzroy's storm 'cones' that are hoisted at harbors to warn shipping. In Britain where the Government had in fact only set up an office to collect meteorological statistics of past weather, forecasts were also issued, but they were unauthorized. It was not long before the scientific community and the Government criticized this practice to the extent that public forecasts were actually suppressed. By popular demand they were reintroduced 11 years later in 1877, but not before Fitzroy had committed suicide in 1866.

Forecasting of weather and, more generally, of the environment continues to be criticized and misunderstood, especially when any new developments are introduced, as one has observed when long-range and seasonal weather forecasts and climate change predictions began, or when environmental forecasts are presented in probabilistic terms.

Forecasters and those who manage them seldom take the opportunity to explain the underlying principles and methods of forecasting, perhaps on the one hand to avoid the continued skepticism of the scientific community, or on the other hand because most of their clients are reasonably satisfied.

This paper is a contribution to opening up some of the issues involved which need more discussion and understanding, as forecasting methods are developed for ever longer chains of environmental phenomena connected to social and economic activities. An important example is the coupling of weather and traffic forecasts to provide the input to calculations of atmospheric chemistry, which lead to forecasts of zone levels; these are then disseminated to the public and urban planning bodies, who then have the information of taking appropriate action.

It is generally acknowledged that as economies become more advanced, the productivity and profitability of many activities become increasingly dependent on environmental information and forecasts, for example, in reducing excess materials and travel times in construction and transportation, and in farming, an increasingly high technology business that continues to be very weather sensitive.

Scientific foundation of the first kind of 'deterministic' forecast

It is essential to distinguish between two different kinds of environmental forecasting, which broadly correspond to shorter and longer time intervals (denoted by ?tf) between the forecast and the event. The first kind of forecast, to use the terminology of Lorenz [46], is deterministic or causal in its reasoning and in ...