NCL Home > Documentation > Functions > General applied math

simpne

Integrates a sequence of unequally or equally spaced points using Simpson's three-point formula.

Prototype

	function simpne (
		x  : numeric,  
		y  : numeric   
	)

	return_val  :  float or double

Arguments

x

An array of one or more dimensions. The rightmost dimension is the dimension to be used for the integration. It must be strictly monotonically increasing. If y is multi-dimensional and x is one-dimensional then x must be the same size as the rightmost dimension of y. Otherwise, x and y must be the same size.

y

An array of one or more dimensions. The rightmost dimension is the dimension to be integrated. This may require that the original variable be reordered. Given that Simpson's 3-point formula is used, there must be 3 or more non-missing values present.

Return value

The return value will have same dimensions as all but the rightmost dimension of y (or be a scalar if y is one-dimensional). The return type will be double if y is double, and float otherwise.

Description

This function uses Simpson's 3-point formula for integration adapted for unequally sampled points. The number of points may be odd or even. The more points (finer resolution) the more accurate the answer. The x must be strictly monotonically increasing. Missing values are allowed. They are ignored.

See Also

ftcurvi, simpeq, wgt_vertical_n

Examples

The more points (finer resolution) the more accurate the answer.

Example 1 Evaluate exp(x) from x=1.8 to x=3.4.

 
   x  = (/1.8, 2.0, 2.33, 2.49, 3.122, 3.4/)    ; even number pts (6)
   y  = exp(x)                   
   simpne_6 = simpne(x,y)       ; = 23.9757     ; even number pts (6)

   X  = (/1.8, 2.0, 2.33, 2.49, 2.75, 3.122, 3.4/)
   Y  = exp(X)                   
   simpne_7 = simpne(X,Y)       ; = 23.911      ; odd number pts (7)

The use of a function (here, exp) is for convenience. The fi (here, y) could be from a table or series of unequally spaced values.

Example 2: Assume x is dimensioned ntim and y is dimensioned nloc x ntim. Integrate over time at each location:

   result  = simpne(x,y)         ; result(nloc)

Example 3

Assume z is dimensioned lev and f is dimensioned ntim x klev x nlat x mlon with named dimensions "time", "lev", "lat", and "lon". Integrate over z:

  result  = simpne(z,f(time|:,lat|:,lon|:,lev|:)) ; result(ntim,nlat,mlon)

Example 4

The values of x must be strictly monotonically increasing. Specifically, this means that every x(i+1).gt.x(i). Consider x[*] and y[*]:

   if (isMonotonic(x).eq.1) then  ; =1 means strictly monotonic
       result = simpne(x,y)     
   else
       nx     = dimsizes(x)
       imono  = ind((x(1:)-x(0:nx-2)).gt.0)
       result = simpne(x(imono),y(imono))
   end if
   

Contact the Webmaster