Hi Saurabh,
You have units of pressure in hectopascals not Pascals, therefore you are out by a factor of 100.
Cheers,
Russ
On 13/06/18 09:55, saurabh rathore wrote:
Dear Satyesh,
Thank you for your profound email. But my query is still no resolved and I having confusion for the calculation of vertically integrated moisture flux and its divergence over the data levels of (1000, 925, 800, 700, 600, 500, 400, 300). I have 3 results for divergence of vertically integrated moisture flux that I computed using NCEP and ERA-Interim and according to my understanding its unit is (kg/m^2/s) which is equivalent to (mm/day). Then I compared this computed result form ERA-Interim provided data for divergence of vertically integrated moisture flux and I found that my computed result is 2 order less than this.
I am attaching figures and my script for this computation. I am trying hard to understand why it is so but absolutely clueless about this.
regards, saurabh
On Tue, Jun 12, 2018 at 11:58 PM, Satyesh Ghetiya <satyeshghetiya@xxxxxxxxx> wrote:
India.G Satyesh,Regards,I hope this helps.1*(2-0)+1*(5-2)+1*(10-5)= 10Things will look systematic in regular grid, but in case you have irregular grid (which I presume will be in your pressure axis case) then each box size (dP) will vary for each level of integrand.You try different limits for @din and check its calculations and match with your requirement. Also while using 'list variable[k=1:N@din]' ferret will tell you which range it has taken for the calculation. Also look @din function in ferret manual and how it is taking the point variable's value in which box for integration calculation.Hi Saurabh,It looks okay to me, but it is very important to convince yourself what is going on in ferret. First look into https://www.pmel.noaa.gov/mail
lists/tmap/ferret_users/fu_201 and then follow.4/msg00216.html
By small illustration it will be clear that how ferret does the calculation :
Consider below, the irregular grid which is at 1m, 3m, 7m and 13 m levels. Consider a variable named 'one' which has value 1 at these points till 7m, after which it has no value(...) at 13m grid point.
*****1***** 1
*****3***** 1
*****7***** 1
****13***** ......
So if you do " list one[k=@din] ", it will give:
In this calculation, '2' is midpoint of 1 & 3. Subtract (3-'2') from 1 to get the previous midpoint, which is 1-1=0. so first box is 0 to 2. Then 5 is mid point of 3 & 7, so 2nd box becomes 2 to 5. Then midpoint of 7 & 13 is 10, so 3rd box is 5 to 10. So likewise even if we give one[z=1:3@din], result will be 10 which considers the next level (13 also) indirectly in calculation, but only the depth level, not the value at that level. Observe that 1 is multiplied with each box size, which is the value of the variable.
On Mon, Jun 11, 2018 at 7:04 AM, saurabh rathore <rohitsrb2020@xxxxxxxxx> wrote:
Dear Ferreters,
I am having a small doubt regarding vertical integration for the computation of vertical integrated moisture flux and the vertical axis is having pressure units.
This is the required formulation for this computation
Moisture flux = (1/g) * integration (0 to surface pressure) of [ specific humidity * horizontal wind vectors ] * dP.
In my data the vertical axis is in hectaPascal from 300 hPa at k=1 to 1000 hPa at k=8. So if I integrate [ specific humidity * horizontal wind vectors ] vertically, it will integrate from 300 to 1000 hPa, so is it doing this integration using correct dP weights i.e thickness of the pressure levels ?
The command I used for this computation is
let zon=(qm[g=u]*u*(1/9.81))
let mer=(qm[g=u]*v*(1/9.81))
let uvec=zon[k=@din]
let vvec=mer[k=@din]
is it correct ?
regards, saurabh
--
REGARDS
Saurabh RathoreResearch Scholar (PhD.)Centre For Oceans, Rivers, Atmosphere & Land Science TechnologyIndian Institute Of Technology, Kharagpurcontact :- 91- 8345984434
--
REGARDS
Saurabh RathoreResearch Scholar (PhD.)Centre For Oceans, Rivers, Atmosphere & Land Science TechnologyIndian Institute Of Technology, Kharagpurcontact :- 91- 8345984434