please see this paper:

"Runoff Estimation in Steep Slope Watershed

with Standard and Slope-Adjusted Curve

Number Methods"

and this discussion:

https://www.linkedin.com/groups/effect-terrain-slope-on-SCS-82339.S.193604316

If you have hydro-metric station in outlet  of watershed, you don't need use the CN Method or another for runoff estimation, but if you don't have any hydro-metric station, the CN method is useful and the Rational method is suitable if your catchment is less than 10 km2.  

Most of our technology is based on the Water Erosion Prediction Project (WEPP) model, which uses the Green and Ampt infiltration approach. The model seems to d a reasonalble job of predicting daily runoff amounts, either from a single storm, or we prefer, from a continuous run of 50 or 100 years. WEPP needs improvement, however in predicting peak runoff amounts from watersheds. To do this, we use an NRCS regression equation that predicts peak flow from daily runoff, time of concentration, and curve number, based on the NRCS TR-55 approach. We have an online interface to do this (htt://forest.moscowfs.wsu.edu/fswepp). On this link, select the Peak Flow Calculator. Let me know if you want more informatoin on WEPP or the Peak Flow Calculator.

SWAT model have been tested in different watersheds. I have used the model for different watersheds and I found it very successful. please read some of my articles I published for watersheds in Africa and the Caribbean.

thanks all for giving me so useful information..... 

@William yes any information on WEPP is welcomed, i  read about this model but i have a question regarding the green ampt infiltration approach for runoff calculation... will it be robust...

@setegn... I too used SWAT for my Master's project, if it works well in that environment then nothing can be better than this...

Thanks a lot.....

Bhumika, I hae attached two documents with information about estiamting peak flow rates. One is a spread sheet that compares peak flow predictions using two different curve number tools and two different WEPP watershed tools, and the other is the draft documentaiton for our online Peak Flow calculator.

Note that both the Curve Number, and the time of concentration are important variables in steep watersheds when using the Curve Number, especially if the watersheds are forested.

The Wildcat 5 Curve Number tool is currently in press. It is unlikely to be available in time for your research. When it becomes available, I will send it to you.

Good luck with your MS studies.

@ William thanks a lot sir...

answer of first question is yes

I have done my masters in 2006 and used scs method for finding out runoff and developed a simplified hydrological model,

this has three vertical tanks, surface reservoir,soil reservoir and ground reservoir

the area selected is of hilly region,river of 40 km only.

I fully agree with the previous comments and I suggest this paper on the application of WEPP to a small Mediterranean catchment for runoff and erosion predetermination.  Runoff and water erosion modelling using WEPP on a Mediterranean cultivated catchment D. Raclot & J. Albergel Physics and Chemistry of the Earth, Parts A/B/C

Volume 31, Issue 17, 2006, Pages 1038–1047

Thanks sir..

All the studies we did in Sri Lanka and even in Thailand did not give good results with the SCS curve number method. The Runoff and its peak is not predicted correctly.  Please refer the following publications of ours. Good methods to estimate runoff are given in the paper number 1 given below.

1.Halwatura, D., Najim, M.M.M., 2013. Application of the HEC-HMS model for runoff simulation in a tropical catchment, Environmental Modelling & Software, 46 : 155-162.

http://dx.doi.org/10.1016/j.envsoft.2013.03.006

2. Najim. M. M. M. Babel M. S. and R. Loof. 2006 AGNPS Model Assessment for a Mixed Forested Watershed in Thailand, ScienceAsia 32(1): 53-61

3. Babel, M.S. Najim, M.M.M. and Loof, R. 2004. Assessment of AGNPS Model for a Watershed in Tropical Environment, Journal of Environmental Engineering of the ASCE, 130(9): 1032 – 1041.

4. D. Halwatura, Najim M.M.M. and N.W.B.A.L Udayanga, 2013. Impacts of Urban Water Management in Attanagalu Oya Basin, In proceeding of: Special session on Urban Water Environment Monitoring & Management, 4th International Conference on Structural Engineering and Construction Management, 2013,, At Kandy Sri Lanka.

SCS curve no method we can use for hilly terrain also. I think there is a chart for landuse basis curve no. You can go for it.

This discussion may be lacking a careful delineation of the NRCS (SCS) curve number method and the SCS method of estimating peak runoff. As Hawkins et al. 2009 (Curve Number Hydrology, ASCE Press, Reston, Virginia) notes, the CN method estimates runoff volume. The SCS peak flow estimate is based on the CN but these two methods are kept separate in the NRCS National Engineering Handbook. 

The CN conceptual model (Initial abstraction, retention, and runoff)  supplemented with algorithms for evapotranspiration and various moisture storage or retention processes is used in 60% of the models of watershed hydrology. See Negussie Tedela, Dissertation, Warnell School of Forestry and Natural Resources, University of Georgia, Athens 2009. These models are widely used in mountainous terrain as several noted in this discussion. However, these watershed models all require calibration and thus requires gaging data for rainfall and stream flow. 

The CN method was designed specifically for ungaged watersheds. Most but not all of the watershed data available in 1954 (24 watersheds) were from flatter cultivated areas in the US. However, some catchments were on range lands in the western mountainous US. At the very least, hilly, piedmont watersheds were represented in 1954. Steeper terrain was covered in supplemental arid lands curve number tables involving shrubs, bushes, and sparse tree cover in the western US dating to ca. 1960s.

Despite the paper recommended by Noor, and 1-2 papers published in other lesser journals including some work in China, I do not find a scientifically sustainable relationship between slope and the watershed curve number. 

See Tedela, N. H., S. C. McCutcheon, T. C. Rasmussen, C. R. Jackson, E. W. Tollner, W. R. Swank, R. H. Hawkins, J. L. Campbell, and M. B. Adams. 2012. Runoff curve numbers for 10 small forested watersheds in the mountains of the eastern United States. Journal of Hydrologic Engineering 17(11): 1188-1198. (online date) doi: 10.1061/(ASCE)HE.1943-5584.0000436 [http://ascelibrary.org/doi/abs/10.1061/%28ASCE%29HE.1943-5584.0000436]

Furthermore, I would not expect that the fundamental CN concept (initial abstraction, retention, and runoff) would be invalidated on steeper mountainous slopes. If we could distinguish any slope effects from the significant background uncertainty dominated primarily by event volume effects (Hawkins 1993 J Irrigation and Drainage Engr and Tedela et al. 2012 noted above) and effects of forest cover on runoff, I would expect less uncertainty in the CN relationship the steeper the slopes. Fundamentally, anything that speeds up the runoff response to rainfall including reducing any storage effects, will improve the applicability of the CN method.

That said, steeper mountainous terrain has  at least two complications that erratically delay quick runoff responses--snow accumulation and forest litter. The NRCS National Engr Handbook explicitly rules out application of the method to snowmelt (yet at least one group has derived very specific curve numbers for seasonal snowfall and snow melt for at least one watershed). The paper above shows that 90% of the mountainous forested watersheds were not amenable to CN estimates using the tabulated NRCS CN for woodlands.

Wm Elliot and most others submitting comments presume that peak runoff, rather than volume will be of greater interest in addressing this question. Elliot remarkably provides insight into work underway at the renowned USFS Intermountain Mt facility. Hawkins and other CN experts commend the Wildcat software. WEPP has been in development since at the least the 1970s. Ms. Uniyal might be well served to volunteer as a reviewer or beta-tester of Wildcat 5.

That said, Elliot does not mention alternative software, particularly the TOP Model by Blevins of the UK (primarily). Tedela 2009 evaluated the TOP Model for four of the ten watersheds used to evaluate the CN method with calibration and found that the most fundamentally sound watershed model could only estimate runoff from 3/4 watersheds. This was a much more rigorous test than SWAT and other highly parameterized models have ever undergone. Thus we do not have many reliable options for ungaged watersheds in mountainous terrain.

Finally, one contributor recommends use of the CN method or the Rational Method. Having taught hydrology in 3 US universities, I consistently recommend my students not use the rational method in practice, unless a particular city or other engineer in charge may have extensively calibrated the method in small watersheds.  My perception is that the fundamental derivation of the rational method is even weaker than that of the CN method, a method that is beset by significant uncertainty. 

I join Elliot in wishing you well in your studies.

Yes CN method is most popular but I think the spatially distributed model based on DEM (digital elevation model) would be more suitable as they calculate more accurately the slope in the mountianous terrian. The slope is the major driving force for runoff, therefore I would suggest any of the MIKE model for this purpose. 

Hi Sakshee Rana,

You may try with the SCS CN for your watershed only for event based simulations. Please give a try. It might work.

There are several adaptations of CN method that take in account the effect of slope gradient on runoff volume.

In general terms, simulation models deal with runoff under two different approaches: Curve Number and/or Green-Ampt. There are several adaptations for each method for considering slope effects on runoff (see EPIC, APEX, SWAT, WEPP documentation). And, it is must be said, even within the Green-Ampt approach, CN is applied to resolve some variables or parameters (see WEPP).

SWAT model (based on CN and consequent adaptations for considering slope and soil moist effects over runoff) after calibration, performed good monthly simulations of water yield at mountainous watersheds in Venezuela (even though daily simulations were not so good).

I think that a interesting and practical approach would be to apply CN method considering the slope effect on runoff as EPIC, APEX and SWAt do. These equations are available in the documentation of each model.

For peak flow estimation, it is necessary to consider the effects of soil (properties and moisture), land use and slope on concetration time calculations. For example, Kirpich method is totaly blind to these variables. It is possible that inaccurate peak flow estimations have been a result of a non appropiate concentration time method and not because CN method by itself.

There are many studies that recommend the application of the SCS curve number method for runoff estimation after proper adjustment of the CN values for slope. If someone (@Bhumika Uniyal) is still interested then I can provide references of different articles.