Re: [ferret_users] Number of cells

["JoshAB1995@xxxxxxxxxxx" <joshab1995@xxxxxxxxxxx>]

Hi both,

Thanks again, however still getting a funny result. And i dont think rsum is very suitable here - the trend in the data should be positive rather than the cumulative number of cells.

Thought you might have been onto something karl as it is possible that some of my values become 0 through the time period, but the changes were made to no avail. 

Thanks anyway, 

Josh

On 20 Mar 2019 04:14, Ryo Furue <furue@xxxxxxxxxx> wrote:

On Wed, Mar 20, 2019 at 3:28 AM JoshAB1995@xxxxxxxxxxx <JoshAB1995@xxxxxxxxxxx> wrote:

Dear Ferreters,

 

I’ve been stuck for a considerable length of time trying to calculate the number of grid cells that contain a desired value, which I could then multiply by some value to gain a volume.

 

For the values I’m after I’ve created the following mask:

 

Let X = if x LT x then 1 else 0 !This sets all of the values that are within the range I’m after as ‘1’

Set var/bad = 0 X !To remove the values from my dataset that fall outside of the desired range

Plot X[x=@ngd,y=@ngd,k=@ngd] !Hoping that this plots the total number of grid cells that are not ‘missing’ against time

 

From this plot I get an unexpected result (a decrease when it should be increasing!!!)

But you are plotting "the number of missing cells at each time step".  That doesn't have to increase.  Perhaps what you want is

let cumulative_sum = X[x=@ngd, . . . , L=@rsum]

I guess.  @rsum is short of "running sum":  rsum(L) = sum from L' = 1 to L' = L of X(L').

Cheers,

Ryo