hurricane, their analysis is necessarily rather subjective. Some forecasters claim dependable results in using this concept when data were available to high levels near the storm. The technique is not based on the assumption that wind at any single level is responsible for steering the storm, since the forces controlling movement are active   through   a   deep   layer   of   the   atmosphere. However,  as  successive  levels  are  analyzed,  a  level  is found at which the closed cyclonic circulation of the storm   virtually   disappears.   This   steering  level coincides with the top of the warm vortex and varies in height with different stages and intensities of the storm. It maybe located as low as 20,000 feet or, in the case of a large mature storm, as high as 50,000 feet. It has been found in analysis that most weight should be given to the winds in advance of the storm within a radius of 200 to 300 miles in preference to those in the rear quadrants. The hurricane generally moves with a speed of 60 to 80 percent of the current at the steering level. THE  DIRECTION  OF  MOVEMENT.—   The direction of movement is not always exactly parallel to the steering current, but has a component toward high pressure that varies inversely with the speed of the current, ranging from almost 0° with rapid movement to as much as 20° with speeds under 20 knots. In westward moving storms, a component of motion toward high pressure could result from the poleward acceleration arising from the variation of the Coriolis parameter across the width of the storm. This would indicate that, to the extent that this effect accounts for the component of motion toward high pressure, northward moving storms would fit the direction of the steering current more  closely  than  westward  moving  ones.  The tendency for poleward drift would be added to the speed of forward motion in the case of a northward moving storm so that it would approach more closely the speed of the steering current. Empirical evidence supports this hypthesis. Corrections for both direction and rate of movement should  be  made  when  this  is  indicated  by  the windflow downstream in the region into which the storm  will  be  moving.  For  prediction  beyond several  hours,  changes  in  the  flow  pattern  for  a considerable  distance  from  the  storm  must  be anticipated. It  should  also  be  remembered  that intensification or decay of a storm may call for use of a higher or lower level, respectively, to estimate the future steering current. THERMAL STEERING.— A number of efforts have been made to correlate hurricane movement with thermal patterns. For example, one writer suggests that a storm will move along tongues of warm air in the layer from 700 to 500 hPa that often extends out to 1,000 miles ahead of the storm. The orientation of the axis of the tongue then may be regarded as a reliable indicator of storm movement for the next 24 hours. A considerable difficulty in applying this technique is created by the fact that the warm tongue sometimes has more than one branch and it is questionable as to which is the major axis. RECURVATURE One of the fundamental problems of forecasting the movement of tropical cyclones is that of recurvature. Will the cyclone move along a relatively straight line until it dissipates, or will it follow a track that curves poleward  and  eastward?  When  recurvature  is  expected, the forecaster must next decide where and when it will take place. Then, he or she is faced with the problem of forecasting  the  radius  of  the  curved  track.  Even  after the cyclone has begun to recurve, there are a great variety of paths that it may take. At any point, it may change  course  sharply. The  most  common  recurvature  situation  arises when an extratropical trough approaches a storm from the west or when the storm moves west to northwest toward a stationary or slowly moving trough. Indicators of Possible Recurvature Some of the indicators of possible recurvature are as  follows: .  If  the  base  of  the  polar  westerlies  lowers  to 15,000 to 20,000 feet west of the storm’s latitude, and remains  in  this  position,  recurvature  may  then  be expected  to  occur. . However, if there is the building of a dynamic high or an eastward movement of this high to the rear of the advancing trough, and the westerlies dissipate in the low latitudes, the storm will move past the trough to its south and continue its westward path. l The above rule also holds true in cases where the polar trough moves from the west against a blocking high. The higher latitude portion of the trough continues to move eastward while the southern segment of the trough is retarded and is no longer connected with the upper portion of the trough. l Recurvature may be expected when an anchor trough is about 500 miles west of the storm and when the forward edge of the westerlies is from 500 to 700 11-6