There are quite numerous examples where US Geological Survey has used available streamflow data within each physiographic area and regional or state zones to help estimate floods. However, they have not used their data in estimating un streamflow on an ongoing basis. Storms move across a landscape or region, so the timing of flows, hydrographs, etc. are not synchronized to ungauged areas. So then, it is probably best to select a model, validate to your available rainfall, streamflow, physiographic area, soils, land use, irrigation, well levels, etc. data, and use it to help predict. To some degree, predicting streamflow requires more information and effort than actually instrumenting and measuring water with vented transducer, and using a standard flume, weir or developing stage discharge rating curves. To some degree, surface streamflow can also be impacted by channel substrate materials and whether stream is loosing or gaining flow from water table from adjacent areas. To some degree, you might find indicators in aquatic insects, plants, mosses, water stain lines to assess whether stream is perennial, intermittent or ephemeral. It may be best to interact with hydrologists and civil engineers in the region of interest to see what they use and have confirmed.

I recommend: Article Quantificação de vazão em pequenas bacias sem dados

Dear Mamuye,

We find that it is often best not to downscale or regionalise your reconstructions and predictions of streamflow directly because large scale factors can reduce noise and give better estimates the site specific downscaling. See for example this study:

https://iopscience.iop.org/article/10.1088/1748-9326/10/6/064006/meta

where we found that using large scale atmospheric indicators gives better results than site specific modelling.

Best regards

Adam

Thank you all for your answer and recommendation.

Dear Mamuye Tebebal,

In one single question you have just captured one of the most challenging (yet still unresolved) question in hydrology, that is, Prediction in Ungaged Basins (PUB).

You might want to take an indepth look at the work the the IAHS working group (https://iahs.info/Commissions--W-Groups/Working-Groups/Prediction-in-Ungauged-Basins.do) on the matter (https://iahs.info/pub/pubs.php) as it advocates the application of traditional ways to predict discharges in ungauged catchments (that is area proximity, regression analysis,...).

Article IAHS decade on Predictions in Ungauged Basins (PUB), 2003–20...

Article A decade of Predictions in Ungauged Basins (PUB)—a review

Dear Roland Yonaba,

Thank you very much for your interesting suggestion

How can I say.. Depends. We have a lot of technics that can be applied to estimate flows data in ungauged basins. We can use regionalization based on area, precipitation and other kind of data. But to use this approach is necessary to have a gauged basin with a good temporal flow series next to your ungauged basin. We can use rainfall-runoff technics too. But this approach, in the most cases, are more difficult to be applied. Depending on your area, try to use models as HEC-HMS or MGB IPH models.