You can use synthetic unit hydrograph of similar watershed or some lumped model for which coefficients are available for your area. Or search if some body have calibrated model in your area.

If rainfall and land cover are similar in the ungauged catchment as the larger catchment containing the ungauged catchment, a simple approach would be to take a ratio of the areas and use that to figure out what fraction of large catchment discharge is the smaller ungauged catchment discharge. However, this requires a similar spatiotemporal rainfall, similar soil properties and land cover (similar vegetation types, farms, urban), and no major differences in water abstractions (like surface irrigation from a river)  between the two catchments. While one can do anything with models, one should be aware of the errors that can arise, if the hydrological inputs and processes differ between the ungauged and gauged catchments..

if you are designing for an important and costly project and if you have the time and resources, it is probably worthwhile to do a short series measurements of your ungauged catchment for comparison with the long time series of your gauged catchment. this would give you very valuable information for assessing how close to reality is your model outcome.

Hello Philip, Prediction in the ungauged  (PUB) basin is the most challenging problem tempting the hydrologist. However, there are traditional ways to predict discharges in the ungauged catchments (e.g. area proximity, regression analysis etc). Thanks to the PUB initiative  in the last decade (2003-2012) many efforts have been done to resolve this issue. There are many research findings that have been published see e.g the summary by Hrachowitz, et al., 2013 A decade of Predictions in Ungauged Basins (PUB)—a review. Hydrological Sciences Journal, 58(6): 1–58. Further more see also  Montanari et al., 2013. The continued initiate to the PUB, the Panta Rhei—Everything Flows”: Change in hydrology and society—The IAHS Scientific Decade 2013–2022.

I think in your case, depends on your purpose you can estimate the discharge in the ungauged part using SWAT model. Since SWAT model can account the spatial heterogeneity and at HRU in each sub-catchments, there is a possibility to estimate the discharge at desired sub catchments. Nevertheless, you have to be careful about the uncertainity associated with both the inputs, model structure and parameters, which influence your output. I think it helps you a bit to start your estimation.

Dear Philip, you've mentioned that you have only runoff (discharge) data for larger catchment in which your catchment is nested. It is not mentioned if at least some precipitation data are available for your purpose. If not, I would recommend you the approach considering the ratio of areas mentioned by Amartya. However, you should be aware of the fact that the error will be generally increasing with increasing difference of areas due to different reasons such as different spatial distribution of precipitation, different geology and many others mentioned also here in discussion. Another case would be if you have available precipitation data of satisfactory spatial and temporal distribution. In this case, you can use the model to calculate runoff from your catchment. I would also like to mention another issue which is the purpose of model application. You should do carefully the calibration of your model with respect to the purpose. As you mentioned that you are focusing on water supply, the calibration should put less emphasis to flood flows and more to usual situations and low flows. 

you can calibrate the gauged catchment using SWAT model and at the same time you have to assign an outlet to the ungauged watershed (as you have said it is located inside the gauged catchment), then using SWAT you can calibrate the gauged flow, at the same time based on your calibration you will have outflow for the ungauged part.

You can also make area proportion this is rough assumption actually.  

I agree with Gerald.

In addition, if you have data and time, in my opinion you can apply  the SWAT model at  the larger basin.  As your ungauged catchment is also located in this larger catchment, you will set the the ungauged basin as a subbasin of the larger basin.

However, for catchments with few or no discharge measurements, Bárdossy A. (Calibration of hydrological model parameters for ungauged catchments. Hydrology and Earth System Sciences 11, 703-710. 2007) assumes that catchments with “similar” characteristics show a “similar” hydrological behavior and thus a transposition of model parameters from “similar” basins is plausible. In this case you have to identify a similar basin.

In alternative, you can try to identify a relation between the  river discharge and the physical and hydrological characteristics of the corresponding drainage area of the larger basin and to operate a translation in your basin. However, this is quite difficult and the results are affected by a very large uncertainty. You have to analyze deeply climate, land use, slope, soil type in order to identify valid relationships. 

I believe WARMF can also do the job for you like SWAT. You can Google WARMF and download it from EPA website. You can apply WARMF to your large watershed and calibrate it with available gauged data. Your un-gauged smaller watershed will be included within the boundary of entire tributary land. You can output the outflow for the terminal point of the smaller watershed with WARMF. 

I suggest you this simple but highly efficient method, named Rating Curve Model, developed by my colleagues. Attached you can find the last paper describing the method.

Article Enhancement and comprehensive evaluation of the Rating Curve...

Hello Colleagues. thank you for all the answers. Like Sirak said, prediction in ungauged catchments is indeed very challenging. I will try out all the options suggested and choose the most convenient.

The rating curve model proposed by Barbetta is new to me but seems interesting.

But generally, I feel that many models donot take into account the component of base flow, which can completely change the hydrograph behavior especially during low flows.

Thanks

'Price pass through effect within Uganda's cotton industry following trade liberalization' American Journal of Economics, 4(4): 159-170.

Dear Philip

Large number of studies have been done in past. Please go through the critical papers, few of them I am attaching, hope it works for you. 

http://essay.utwente.nl/57578/

http://www.hindawi.com/journals/geography/2013/214140/

https://www.google.co.in/url?sa=t&rct=j&q=&esrc=s&source=web&cd=9&cad=rja&uact=8&ved=0ahUKEwjLlt7xyKHOAhXKu48KHfnWDSsQFghmMAg&url=http%3A%2F%2Fwww.uni-stuttgart.de%2Fgi%2Fresearch%2FGeodaetische_Woche%2F2013%2Fsession05%2FSneeuw-etal.pdf&usg=AFQjCNFBaq64qe7HTcv4Frrh7oPYyRwQnw&sig2=ZYquCD3qYNTf5UaYfZrgfw

If need any further help, please feel free to contact.

Its a big issue. It necessarily needs to be region specific  either on paired catchment concept or statistical concepts, but without forgetting the physics involved therein.. 

Hi,

How about if flow data for larger catchment is unknown but smaller catchment that lie within larger catchment is known?