Double mass curve: which line segment is extended to find adjusted slope?

  • Thread starter tzx9633
  • Start date
  • Tags Tags
    Curve Mass
9 replies · 5K views · last reply
tzx9633

Homework Statement


In this first photo , we can see that the adjusted is found by extending the longer portion of straight line . The line is extended to the upper part of the graph.( We can see that there's a break in the line , the longer portion of the line is extended , and taken as adjusted slope) . And the shorter part of the line is taken as the original slope .
So , the lower part is the adjusted slope .

But , in the 2nd example , we can see that the adjusted slope is found by extending the loner portion of line too . But , in this , the line is extended to the below part . And the shorter portion of line is taken as original slope...

P/ s : Photo 707 and 708 are for case 1 , the others are for case 2 .

Homework Equations

The Attempt at a Solution


So , based on understanding , can I conclude that for double mass curve . there's always a break in the line . So , for the shorter part , it's always taken as the oriiginal slope . For the longer part , it's extended to get the adjusted slope . Is my concept wrong ?
 
Refer to here
 
Attachments
  • Numbered list of five steps for a double mass curve analysis of rainfall data: selecting base stations, cumulating their annual averages, plotting against station X, identifying a change in slope, and adjusting inconsistent data.
    707.PNG
    32 KB · Views: 793
  • Double mass curve graph of cumulative annual rainfall at station X against cumulative average annual rainfall of all base stations, with a straight line through the origin and a steeper line, showing the deltas ΔPo and ΔPa and slopes Mo and Ma, beside the equations Pa = Po(Ma/Mo), Mo = ΔPo/ΔP and Ma = ΔPa/ΔP.
    708.PNG
    17.7 KB · Views: 967
  • Printed problem text on precipitation gauges: annual precipitation records for five stations A to E from 1926 to 1942, with gauge A relocated after 5 years and records to be adjusted for 1926 to 1930 using gauges B, C, D and E.
    709.PNG
    16.3 KB · Views: 881
  • Table of yearly rainfall data from 1926 to 1942 with columns for annual rainfall at A, annual rainfall at stations B, C, D and E, average rainfall of base stations, and cumulative rainfall for A and B+C+D+E.
    710.PNG
    20.2 KB · Views: 826
  • Double mass curve plotting cumulative precipitation for Station A in mm against cumulative precipitation for Stations BCDE in mm, with year labels 1926 to 1942 along a rising line and two red-circled ratio annotations.
    711.PNG
    10.7 KB · Views: 893
  • Double mass curve plotting cumulative precipitation for Station A in mm against cumulative precipitation for Stations BCDE in mm, with a straight line of points labelled by year from 1926 to 1942; two regions are circled in red showing ratios 115/620 with M0 = 0.19 and 531/2011 with M0 = 0.26.
    711.PNG
    10.6 KB · Views: 1,199
  • Handwritten calculation of adjusted precipitation values: Pa = Po |Ma/Mo|, then P1926 = 32.9 |0.26/0.19| = 32.9 (1.37) = 45 mm, P1927 = 28.1 (1.37) = 38.5 mm, P1928 = 33.5 (1.37) = 45.09 mm, P1929 = 29.6 (1.37) = 40.55 mm.
    712.PNG
    13 KB · Views: 795
Hydrology isn't really my subject but...

I've read the explanation in 707 and it looks like they are assuming that the part of the curve that has more data points is the most accurate. They then extrapolate to "correct" that part of the line for which they have less data. So..

In the first example (708) most of the data is at the lower end so they extrapolate upwards.
Double mass curve plot of cumulative annual rainfall at station X against cumulative average annual rainfall of all base stations, with a red arrow marked less data on the upper segment and a green arrow marked more data on the lower segment, plus equations Pa = Po (Ma/Mo), Mo = delta Po / delta P and Ma = delta Pa / delta P.

In the second example (711) most of the data is at the upper end so the extrapolate downwards.
711.PNG
 
  • Like
Likes   Reactions: tzx9633
tzx9633 said:
can I conclude that for double mass curve . there's always a break in the line .

No.

707 says that a change in slope implies part of the data is "inaccurate or non homogeneous". If you have a data set that is accurate and homogeneous you will just have a straight line with no change in slope. For example suppose in the second case you only had data from 1930 to 1942.
 
  • Like
Likes   Reactions: tzx9633
CWatters said:
Hydrology isn't really my subject but...

I've read the explanation in 707 and it looks like they are assuming that the part of the curve that has more data points is the most accurate. They then extrapolate to "correct" that part of the line for which they have less data. So..

In the first example (708) most of the data is at the lower end so they extrapolate upwards.
View attachment 212168
In the second example (711) most of the data is at the upper end so the extrapolate downwards.
View attachment 212169
so , the conclusion is we extrapolate the site which has less data ?
 
I'm struggling to explain it in words..

If you have two data sets and you want to know if they are correlated you plot one on X and one on Y. If you get a straight line the data is correlated.

For example if you want to prove Y = X2 you can plot a curve of X2 against Y and you should get a straight line. In this case with a slope of 1:1.

If the data isn't correlated you won't get straight line.

If the line isn't straight then you need to decide which part of it is unreliable by looking at the correlation. In the two examples you posted they have concluded that smaller regions of the graph are less well correlated. So they have extrapolated the data from the larger region to the smaller for station X.
 
  • Like
Likes   Reactions: tzx9633
CWatters said:
hey have concluded that smaller regions of the graph are less well correlated. So they have extrapolated the data from the larger region to the smaller for station X.
this theory applies to all type of questions in hydrology when we are asked to use double mass curve method ??
 
CWatters said:
I'm struggling to explain it in words..

If you have two data sets and you want to know if they are correlated you plot one on X and one on Y. If you get a straight line the data is correlated.

For example if you want to prove Y = X2 you can plot a curve of X2 against Y and you should get a straight line. In this case with a slope of 1:1.

If the data isn't correlated you won't get straight line.

If the line isn't straight then you need to decide which part of it is unreliable by looking at the correlation. In the two examples you posted they have concluded that smaller regions of the graph are less well correlated. So they have extrapolated the data from the larger region to the smaller for station X.
Thanks , i really appreciate your effort to explain ... Can you help me in the following thread ?
https://www.physicsforums.com/threads/missing-precipitation-data-hydrology.927607/

This is a hydrology problem .
 
CWatters said:
they have concluded that smaller regions of the graph are less well correlated.
why is it so ? Cant be the larger region of the graph are less well correlated ?
 
I'm afraid I don't know the answer to that.