WORLD WEATHER NETWORK 
By ATHELSTAN F. SPILHAUS 
University of Minnesota 
Introduction 
To solve the problem of covering the earth with a 
meteorological network, international cooperation 
must be presupposed even though the present time 
seems inauspicious. However, meteorology has an 
enviable record of international cooperation which has 
withstood two major wars and has been maintained, 
albeit in a somewhat restricted manner, in times of 
severe international stress. Cooperation started as a 
means of exchanging observations between nations. 
The actual planning of networks has hitherto taken 
place on the basis of individual national needs. Con- 
sequently, an extremely uneven coverage is afforded 
the meteorologist who desires to study world-scale 
problems rather than regional synoptic meteorology. 
Air navigation, which now encompasses the earth, has 
given a new impetus to the mternational organization 
of weather networks. However, even here the planning 
is directed to the practical needs of air operations and 
seldom results in a network satisfactory for the study 
of the general circulation. 
That the need for such studies is well recognized is 
indicated, for example, by the Southern Hemisphere 
map analysis project being conducted at the Massa- 
chusetts Institute of Technology by Willett [8]. Evi- 
dence of extensive consideration of the problem is 
indicated by numerous resolutions of the Conference 
of Directors at the meeting of the International Meteor- 
ological Organization in Washington, D. C., in 1947. 
Resolutions 21, 36, 109, and 210 [6], dealing respectively 
with meteorological reconnaissance in areas with in- 
adequate coverage by other means, with various ways 
of amplifying the information from oceans, with the 
density of land stations, and with the proper utilization 
of suitable islands, point out the necessary steps in the 
direction of establishing a suitable world weather net- 
work. 
Studies such as those being conducted in connection 
with the Southern Hemisphere project will reveal in 
detail the deficiencies in the world network. However, it 
is the purpose of this paper to look at the problem from 
a comprehensive point of view. 
In general, the world weather network is deficient 
in the two polar areas and in the equatorial zone. 
Latitude for latitude, it is more lacking in the Southern 
Hemisphere than in the Northern; for a given latitude, 
it is always more inadequate in ocean regions than in 
land areas. Because of the general distribution of land 
and sea on the earth, the Southern Hemisphere at 
higher latitudes presents a tremendous and very diffi- 
cult problem of weather coverage. 
The Gaps 
To localize the worst spots of meteorological terra 
incognita, so as to form the basis of the very long range 
planning which is needed in establishing the world 
weather network, certain assumptions can be made. 
Apart from the antarctic continent, it is comparatively 
simple, from a technical standpoint, to provide the 
requisite density of land stations, although the present 
closeness on land—particularly in equatorial areas north 
of 10°S—is very far from the 100-150-km spacing 
recommended in Resolution 109 [6]. The greater problem 
lies with the other two-thirds of the earth’s surface 
which is covered by the oceans. Here the assumption is 
made that, for convenience and economy, suitably se- 
lected islands and reefs should be fully utilized before 
the further step is taken of arranging observations 
from the surface of the deep ocean. 
The ideal situation, in which all strategically placed 
islands are suitably used for meteorological stations, is 
far from a reality. Because stations are established 
primarily for national regional synoptic needs. a number 
of islands which are highly important from a world- 
network pomt of view are not fully manned and 
equipped, meteorologically, since they are far removed 
from the sovereign nation which is, therefore, little 
concerned with their weather observations. Examples 
are St. Helena in the Atlantic which has no upper-air 
station and Clipperton in the east equatorial Pacific. 
Nevertheless, the meteorological occupation of some 
very important remote islands has been encouraging 
in the years since World War II (Marion and Amster- 
dam Islands being two recent examples). 
To illustrate graphically the areas remaiming after 
continents, islands, and reefs are properly turned to 
account, Fig. 1 shows a chorometric chart of the world 
ocean with isochors representing lines of equal distance 
from land of any kind at three hundred nautical mile 
spacing (five degrees of latitude). The base map used 
for this figure is that devised by the author in 1942 
[11]. The data were revised by F. E. Lukermann, Jr., 
Geography Department, University of Minnesota, from 
a study of the islands missing both from the first choro- 
metric chart published in 1898 [2, 7] and also from the 
later ones for the Atlantic, Indian, and Pacific Oceans 
[9, 10]. Because it is feasible to build structures for 
meteorological purposes on reefs which are either awash 
or covered by water less than five fathoms deep, these 
have been regarded in the same sense as islands in 
Fig. 1. Included in this category are reefs such as the 
Virginia Rocks (three or four fathoms) in the Atlantic 
Ocean; Saya de Malha (about two fathoms) in the 
Indian Ocean; and, in the Pacific Ocean, the Maria 
Theresa Reef (awash), the Ernest Legouvé Reef 
705 
