I own pasture fields on the banks of a river in Dorset, England.
In winter (our wet season) the river sometimes floods the surrounding
floodplain, and it would be a great benefit to be able to predict this
flooding the day before, so I can remove my sheep and cattle before
they get distressed.

I have a well about 300 yards from the river, and a raingauge at
my home 2 miles from the river. Each day I record river level, 
watertable level and rainfall, and a graph of these variables
can be seen on http://www.eyemead.com/bm-12.jpg
Is it possible to use this graph to establish the river level
as an approximate function of Watertable level and rainfall and 
time delay? The precise numbers involved over the previous 17 weeks 
are given on:http://www.eyemead.com/bearmead.htm#stour

Regrettably it is beyond my school maths to work out this function
from the experimental data, but I have a "gut feeling" it can be done
by someone with the right expertise. Can you put me in touch with a 
person or an organisation who could work out such a function for me? 

Once I had this function, I think I know enough maths
to be able to insert values of the variables to predict the river 
level. Past experience shows that the river floods when it reaches 
a height of 320 cms (as it very nearly did on 3 December this year,
after 3 intense bouts of rainfall close together). 

I live on top of a hill out of sight of the river and its floodplain!

---

Request for Question Clarification by mathtalk-ga on 02 Jan 2004 23:06 PST

Hi, doxeyman-ga:

My intuition is that the river flooding is primarily driven by
upstream conditions, though of course local rains would contribute to
rising watertables and take up whatever "slack" might be available for
normal runoff.

Have you tried using the government agencies' flood warning services:

http://www.environment-agency.gov.uk/subjects/flood/floodwarning/

Your tax dollars at work & all that, plus they have a broader data
collection scheme to work with, e.g. capable of correlating the
upstream conditions with your local ones.

regards, mathtalk-ga

Hello doxeyman,

I have some experience in this area, and though I can't answer your
question, I just wanted to offer a couple of observations:

--17 weeks just isn't long enough to establish any sort of meaningful
correlation between your variables and the river height -- you'd
probably need a few years worth of data, at least.

--even if you have the data, the level of the river in your
neighborhood probably has more to do with events upstream than right
at your farm.  Sure, the local rainfall and local watertable are part
of the picture, but they may well be fairly minor contributions in the
overall big picture, as far as flooding is concerned.

Just my "two pence" on the topic.  Best of luck.

pafalafa-ga