Daily flow readings may not be that useful in estimating design storm. With the low cost and high usefulness of transducers to collect and log depth readings at frequent intervals, one would wonder if some jump in investment would be worthwhile.

Common, relatively simple methods include SCS CN method and for small drainages and urban areas the Rational Formula Q = CIA. There are numerous papers and guides on internet for use of SCS Method (TP 155 I think) and Rational Method. Since there is a long term gage within the watershed or basin, you might check and verify how well the SCS method compares with available data.

I do prefer the USGS approach, that takes all stream gauges in a physiographic area, and develops equation using primarily drainage area, sometimes also adding urban area as part of equations. I believe most states or physiographic regions in USA have some equations to consider for ungaged drainage estimates. Some other countries may have also developed equations from their stream gauging network. Even with a fairly good network of stream gauges, the 95% confidence limit for where I live is typically + - 40%. Some years back though after looking at various plots of data in my area, the bankfull flow discharge rate (which is a storm peak with typically a return period of 1.5 years) times 10 gave a reasonable estimate of low frequency peak flow. Some researchers with intensive data prefer MIKE SHE model.

It is important when selecting design storm to also consider risk, length of useful life of structure, and impacts of exceedance of design storm. If I was able to accutately select or calculate a 100 year design storm value, for a culvert whose life is 50 years, each year there is a 1% chance of flow exceedance, so over the 50 years, that equates to 50% likelihood of exceedance. Perhaps some design or armoring of culvert fill materials could be added to prevent total failure.

Dr. David Rosgen used 2 times bankfull flow depth at thalweg as an estimate of floodprone depth, probably best taken at riffle section, and this may equate to maybe 50 year flow estimate. Some stream types have easier to read bankfull flow indicators than others.

For long term investments with high risk values, I would recommend considering development of an envelop curve, which takes the known extreme flow of mod rate to long term gauging sites, and plotting drainage area on x axis, and known historic peaks (or prehistoric if suitable indicators exist), and then drawing a curve above all these points. I am not sayijg to necessarily design to this as a standard, but consider that these flows highlight what is possible. I have to go now.

you can read the following article

https://pdfs.semanticscholar.org/ca0c/9fbceeb1c14ac5919f6b5be713aa349ef78f.pdf

Existe la posibilidad de estimar el caudal de diseño de una cuenca no aforada o monitoreada por un periodo suficienteme largo de tiempo aplicando dos métodos analíticos

Para cuenca de reducidas dimensiones es recomendable el uso del método racional el que exige conocer la curva intensidad-duracion-recurrencia, estimar un coeficiente de escorrentía y un tiempo de concentración. Asumiendo que la duración de la tormenta es máxima cuando la duración de la tormenta es igual al tiempo de concentración, para dicho tiempo y la recurrencia de diseño adoptada se determina la intensidad del evento

Luego se multiplica la intensidad por el coeficiente de escurrimiento y se determina el caudal de diseño. Para cuencas de mayor superficie es recomendable utilizar el Método del Servicio de Conservación de Suelos de los Estados Unidos (USDA) que estima lámina escurrida a partir de una precipitación de diseño y la Curva número ,(función de la clase hidrológica del suelo, el uso y la precipitación antecedente). Posteriormente se determina el caudal pico como el producto de la lámina escurrida divido por la suma de la mitad de la duración del evento y el tiempo de retardo. El tiempo al pico es igual al tiempo de concentración y el tiempo base a 2.67 tiempo al pico (Método del Hidrograma Triangular). Métodos más complejos fueron diseñados por HEC de los Estados Unidos (HMS) y otros organismos

Link to many USGS examples of developing equations for estimating peak flows for standard return periods for various regions. Since rainfall is a major consideration, I would suggest reviewing rainfall frequency maps such as TP 40 by National Weather Bureau when applying regional flood frequency equations, and adjust accordingly when topography or other factors contribute to rainfall differences.

Link - https://acwi.gov/hydrology/Frequency/b17c/supplementary-materials/reports.html

Many thanks for your suggestions. William F. Hansen I actually just have daily stream flow in this area and no precipitation records. I need a solution using only the daily streamflows and effective drainage areas to get flows at an ungauged reach of the river.

Thank you.

I ended up using Equation 1 in https://pubs.usgs.gov/sir/2005/5017/pdf/sir20055017.pdf

It is based on the area ratio method which is simple but straightforward with less data requirements.

Thank you all for your great suggestions.