Dear Sirs,

I am Suresh Babu, a research scholar of Anna University, Chennai, India, doing
research in "Rainfall-runoff modeling especially dealing with SCS CN model
modifications".

I am glad to convey that I am at the verge of final stage of my thesis work.
While I did my first draft report of my thesis work, I (and also my supervisor,
Dr. K. Venugopal) feels something better way of presenting the results in my
thesis outputs.

In this regard, I am very much in need of your kind guidelines and helps on the
following queries.

To brief my work, I have used the USDA-ARS watershed database to propose a
modified version of SCS CN model. After data filtering works like P/S >
0.456inches, etc., I got 82 watersheds data(eventwise data) for the final
analysis. Including existing SCS models, Mishra singh model, and my proposed
models, I have totally 8 models for comparitive analysis. I have used Se and
Bias statistics to compare them in calibration as well as validation. And also
"t-test of related samples" were used to arrive at a conclusion on "significant
difference on the statistics" between the models.

Before going for above statistical comparison among 8 models, My supervisor
adviced me to test each model's validity for the given dataset watershedwise
before comparing one model with the other. ie., Acceptable or not acceptable
'Se' values to be tested to findout the watersheds for which each model provides
acceptable results.

For example, he says that, "if R=0.6 for a watershed data by Model 1, it
explains only around 40% of variance. so the Model 1 may not provide acceptable
results in calibration for that watershed, and can be noted down before
comparitive analysis, to strengthen the outcome of the analysis work."

For this purpose, I have searched to get a acceptable range (values) of Se in
various literatures of "SCS model" as well as in "rainfall runoff models". But I
con't get such guidelines so far.

In this regard, I need your valuable guideline. I like to know the range of 'Se'
or 'BiAS', or 'R' or all the above, to arrive at a grouping of watersheds like,
watersheds that has 'Acceptable results' or 'unacceptable results' based on
above statistic values.

My supervisor told that research works in RR modeling beyond 1970's were tested
like the above, and they have suggested some guidelines on the range of
statistics like R, Se, etc., by which the RR(rainfall runoff) model can be
validated as acceptable one or not, for the data set. I will be appreciating if
you suggested any relavent literatures or even contact persons, for the above
purpose. In the mean while, I am also searching for the same.

Here I have attached the Se values in calibration of Model 1 (M1) of 82
watersheds in ascending order of Se. And also listed the statistical summary of
Se values of all 8 models in calibration. From this, I like to know which are
the watersheds having 'acceptable Se' and 'unacceptable Se values'. I need your
general guidelines on this.

Note: Sy-standard dev. of observed runoff; MQobs-mean of the observed runoff;
Se-standard error of estimate.

M1 results:(Note: I have attached the same file as txt with this mail as the
content below may not appear clear while opening your mail box)

WSName Events Sy MQobs Se
(mm) (mm) (mm)

63103 15 4.1717 5.8208 5.4022
63112 15 8.8046 9.4750 5.4581
26863 36 13.9231 11.3704 6.7886
44026 37 5.7396 5.8286 6.8728
69044 37 9.8723 8.9214 6.9044
44028 40 6.9848 7.6665 7.8309
44020 44 6.9160 8.4685 8.2438
44013 20 6.1303 6.6657 8.9103
44023 42 8.9842 8.1746 8.9386
44007 85 13.2696 11.8178 8.9921
62014 49 18.6457 17.1061 9.0655
44017 37 8.6615 8.5647 9.6811
35009 16 9.4949 10.8449 9.8882
26021 21 9.3489 8.9724 9.9401
69032 31 9.5543 8.8249 10.0145
26029 53 10.4941 10.4955 10.0154
26040 63 11.6743 10.7012 10.0470
44022 31 7.1302 9.4522 10.1289
26010 90 12.1198 10.9922 10.1560
44008 82 12.7710 10.8287 10.1806
34007 47 11.1293 10.9431 10.7415
26791 131 9.6373 9.7578 10.7550
44027 43 9.0744 10.6706 10.8222
26028 53 9.4015 10.2357 10.8846
35010 11 10.2301 15.3314 11.2175
44024 31 8.6303 9.6961 11.4916
44011 49 11.1851 10.2582 11.6090
44009 47 13.3417 12.0965 11.7852
26711 76 10.6391 9.3796 12.0609
44010 61 13.1204 10.2265 12.1105
26030 140 12.2821 10.7395 12.1459
26032 20 8.6132 10.1170 12.1717
26027 38 12.4653 12.9954 12.2935
44014 19 5.9044 8.9996 12.3545
26011 34 11.8960 12.1201 12.4057
69036 15 9.8391 14.7994 12.4461
69045 17 11.1282 13.4914 12.4913
17001 95 22.1141 15.3138 12.5310
26014 59 15.3413 10.7842 12.5397
35008 12 10.6630 15.8801 12.7052
26033 49 11.1931 11.0156 12.8056
26012 44 14.6935 12.3195 12.9230
42015 39 12.9607 14.7399 13.0613
17004 52 20.2314 17.3896 13.1019
42037 27 13.8749 10.4562 13.2315
44001 48 14.6134 12.9868 13.5467
34006 42 13.4229 12.7992 13.6294
26034 48 10.6480 10.9249 13.6664
17002 75 21.7297 15.7697 13.9790
26026 67 11.4805 12.3548 14.0573
26024 21 10.4070 14.4901 14.2240
35003 26 17.4413 14.4993 14.3185
35002 22 13.0009 15.2529 14.3235
42024 70 20.1636 18.7520 14.3441
37001 38 11.1929 10.3631 14.4246
42010 59 20.1552 20.1689 14.4568
62002 39 20.1174 14.9706 14.8544
44003 28 10.9215 16.6570 14.8933
26016 13 16.0291 13.8588 15.0315
62010 29 16.5013 19.1275 15.4813
26036 17 14.0037 15.4892 15.5273
42017 51 14.7288 16.8393 15.7214
26017 22 13.8918 15.2938 15.7923
42028 66 17.3761 16.2956 16.0963
44002 29 13.8726 16.5359 16.3388
10001 60 22.8956 16.9995 16.7243
42003 97 19.5771 19.8283 16.8053
37002 61 22.2161 13.1552 16.8301
42002 47 13.9175 20.7454 16.9301
44004 24 18.2172 19.2916 17.1489
17003 11 21.8828 20.8421 17.3285
42006 125 23.5180 17.6680 17.4145
42012 49 19.9997 21.3530 17.7427
42014 30 15.8213 18.9196 17.8966
42038 26 16.4377 14.9219 17.9292
42016 39 18.3126 18.5301 18.2981
42004 36 16.3793 20.0760 18.7581
42011 38 16.5268 18.6352 18.8492
42007 38 17.4312 20.1907 19.2493
42008 41 18.7420 15.3917 19.4417
42036 37 18.5128 18.8480 20.6648
42035 31 24.2388 24.0673 22.6593

Statistical summary of Se in calibration of 8 models:

Events Sy MQobs -------------------Se (mm)----------------------------------

M1 M2 M3 M4 M5 M6 M7 M8

Min 11.00 4.17 5.82 5.40 4.02 2.83 3.05 3.13 2.92 2.97 2.32
Max 140.00 24.24 24.07 22.66 21.73 19.20 22.15 21.40 19.85 16.86 15.44
Mean 44.55 13.67 13.58 13.19 11.84 9.55 11.29 10.49 9.40 8.97 8.69
Median 39.00 13.06 12.99 12.86 11.41 8.92 10.53 9.85 8.92 8.59 8.09
STDV 26.31 4.72 4.14 3.58 3.61 3.28 3.59 3.50 3.21 2.85 3.07
Skew 1.54 0.36 0.32 0.12 0.23 0.55 0.56 0.58 0.62 0.34 0.32

In addition to the above, I wish your guidance and suggestion to present my 8
models comparison in more efficient ways. For your kind information, some of the
above model pair comparison yield contradict results from an expected one. ie.,
for example, it is expected that M6(3 parameter) should perform better than M3(2
parameter) mostly. But Se values comparison of 82 watershed show only 63% of
watersheds showed the above results in calibration. But M6 is structurally
consistent in its form, when compared to M3 model.

I hope that you might be understood my queries. Sir, Kindly write to me if you
require some more detailed results of my study for the above purpose.

I am expecting your valuable responses. I will be very thankful to you all.

With sincere thanks,
Suresh Babu.

"Defeat the Defeat, Before the Defeat defeats You".

P. SURESH BABU
RESEARCH ASSOCIATE
INSTITUTE OF REMOTE SENSING
ANNA UNIVERSITY,CHENNAI-600025.
PH: 22351723 EXTN: 3121

RES:
5/157-A, "CP.GEETHA ILLAM",
1st MAIN ROAD, SARASWATHY NAGAR,
VALLALAR STREET,
OTTERI EXTENSION,
VANDALUR,
CHENNAI-48.
PH:91-44-22751433