Hi Fisseha,
The @AVE transformation in ferret computes a weighted average by SUM(weight_i * value_i)/SUM(weight_i). You could see the weighting used by setting the value at i=k to 1.0 and all others to 0.0 which will give you weight_k/SUM(weight_i). In your case, you could compute the region's average time series and, then, check the weights used in time and compare it to 1/1830 constant weight used in Matlab.
N.B.: When @AVE is applied simultaneously to the X and Y axes, where X and Y are longitude and latitude, respectively, an area-weighted average (weighted by cos(latitude)) is used.
Dear William thank you a lot for your tips.
The data I have is daily concatenated precipitation of 1830 days.
When
I use "list pre[x = x1:x2@ave, y = y1:y2@ave, l=@sum]/1830" gives the
same answer with Matlab but when I use "list
pre[x=x1:x2@ave,y=y1:y2@ave, l=@ave]" , it gives a much smaller value.
Any one who can help with why this is happening would be highly
appreciated as it helps me to know which kinds of analyses are handled
in what why in ferret.
Thank you, Fisseha
--- On Sun, 2/3/13, William S. Kessler <william.s.kessler@xxxxxxxx> wrote:
From: William S. Kessler <william.s.kessler@xxxxxxxx> Subject: Re: [ferret_users] different results from Ferret and Matlab To: "FISSEHA G. BERHANE" <fishcivilmu@xxxxxxxxx> Cc: "Ferret" <ferret_users@xxxxxxxx> Date: Sunday, February 3, 2013, 1:37 AM
On Feb 2, 2013, at 9:23 PM, FISSEHA G. BERHANE wrote:
> Dear Ferreters, > > I have daily precipitation data over a given geographic location. I am getting different mean values of the data in Ferret and in Matlab (5.7mm/day and 7.0 mm/day, respectively). Which is correct, may some one
help me with what is the cause of the difference? > > Thank you, > Fisseha
Are the timesteps regular? Ferret will recognize if they are not, and @ave will correctly weight the time ranges.
Try plotting the time series with each tool (with the same plot ranges and aspect ratio). Are there differences?
With a simple time series like this, you could simply list a small subset of values (say 10). Compute the mean by hand, then (over the same time range) from Ferret and Matlab. Are these 3 identical? If not, can you see the problem looking at a manageable number of points?
BK
|
|