You can use a number of methods. First estimate the amount of rainfall received in each part of the catchment/basin using either Isohyetal or Thiessen Polygon methods. This will given you mean rainfall for the entire catchment/basin. You can also get the percentage amount of rainfall received by each part of the sub-catchment.

In addition to the spatially distributed rainfall calculations,

1. determine the catchment area of each RWHT.

2. use methods for rainfall infiltration calculations such as Green and Ampt equation, (Curve Number,) Horton equation, Philips equation etc. to estimate the amount of overland flow in the catchment of the RWHT.

3. assess the catchment connectivity to get the amount of rainwater at the RWHT by using e.g. spatially distributed hydrological modelling/GIS techniques.

As well as the data processing methods mentioned here, you might want to consider the data collection and telemetry practicalities. You would want to populate the catchment with a well-positioned rain-gauge network collecting data at a useful frequency together with level meters in your RWHTs. If all these devices have telemetry as well as just datalogging you can calculate the rainfall field and relate it to the change in RWHT volume using a real-time or near real-time automated workflow (for example an R script) without having to make multiple visit the fieldsite (this might be a bit of shame if your fieldsite is very beautiful though!).

The density and the distribution of the raingage network can be a limitation. The simplest way of course is to apply the Thiessen or the Isohyete method to define the aerial distribution of rainfall, and to plot this on the RWH techniques (of which you know the area) within the catchment for which you want to know the input of rainfall. An indirect way might be by measuring the runoff of the different RWH techniques of interest, estimate water losses during the runoff process (evaporation, infiltration) and add the losses to the runoff volume to obtain the rainfall dept on the respective RWH technique during the rainfall event. A more accurate answer could be given is the question was phrased a bit more accurate, i.e. what is the size of the catchment, what is the density of the raingage network, is the catchment flat of hilly, what type of RWH technique do you have in mind, etc.,

I am very interested in all your good answers too. I currently work with volunteerly establishing of 20 rain gardens in parcels in a local area for the local water company. I have wondered about the best way to measure the direct uptake from each rain garden during rain. I think your ideas of direct measuring through simple technologies are the best way to get as close to the fact as possible. This setup need as precise data as possible, because the local politicians are going to decide if LID such as rain harvesting is a real alternative to traditional sewer and rain systems. I look forward to follow this thread more. Thank you for posing this question, Ammar, and thank you all for your good answers.

First of all identify the catchment areas that contribute to EACH of the twelve dams. Secondly, calculate the efective rainfall in each of these catchments as described by Mr Benedict Mutua ( Isohyetal or Thiessen Polygon methods). To determine the effective amount of water that came into EACH reservoir during the rainfall, you should have a pre-determined relationship between Water Level and Volume for each of the reservoirs. Then, knowing the water levels BEFORE & AFTER the rainfall for EACH of the reservoirs, you will be able to determine the EFFECTIVE contribution from EACH of the catchments. This methodology has been used successfully in many operating reservoirs.

Considering the 12 small dams, you may demarcate the sub-catchments in the study catchment. This can easily be done using any GIS software. Then identify the rain gauges which contribute mainly to each sub-catchment. The landuse, soil characteristics and soil properties could also be found from different means. Using these inputs you may simulate the flow at each dam site (You may use HEC-HMS 3.5 for this simulation). Considering the capacity of each RWHT, together with a water balance model you may estimate the amounts of water retained by each tank.

You may refer to the following papers to design you study approach.

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

Boers, T. M., De Graaf, M., Feddes, R. A., & Ben-Asher, J. (1986). A linear regression model combined with a soil water balance model to design micro-catchments for water harvesting in arid zones. Agricultural Water Management, 11(3), 187-206.

Jasrotia, A. S., Majhi, A., & Singh, S. (2009). Water balance approach for rainwater harvesting using remote sensing and GIS techniques, Jammu Himalaya, India. Water resources management, 23(14), 3035-3055.

The water balance approach could be designed based on the following papers

Odhiambo, L. O., & Murty, V. V. N. (1996). Modeling water balance components in relation to field layout in lowland paddy fields. I. Model development. Agricultural Water Management, 30(2), 185-199.

Odhiambo, L. O., & Murty, V. V. N. (1996). Modeling water balance components in relation to field layout in lowland paddy fields. II: Model application. Agricultural water management, 30(2), 201-216.

Not all rain fallen becomes rain received/harvested at your dams. It seems you actually want to know the runoff coefficient in each sub-catchment; this is spatially highly variable. While Thiessen and isohyetal methods are practical for dividing your rain fallen, they cannot be equated to rain received or harvestable since you do not know how much of your rain is intercepted (canopy), evaporates from the soil or rock, transpires from plants or percolates deeper and forms groundwater recharge. This really depends on so many factors such as geology, soil, slope & exposition, vegetation, land use, etc. If you are really interested in the quantities arriving at your dam sites, you will have to measure, in addition to rainfall, the runoff. You could also model it, but since you have only a dozen and small dams, it seems that it would be more advisable to measure the flows in all or some selected representative damsite catchments.

As noted above, you will need to estimate the fraction of rainfall that makes it's way into the dams. This can be done using either a simple model calibrated to data (which you will need to collect - as described above), or a method to estimate this fraction from other sources (e.g. SCS curve number) - though you will need to ensure that the method you use is applicable to your study site. Basically, it comes back to a need for data. The easiest way would be to measure the water level in the dams, and convert this to a volume using information about the dam profile (again as noted above). This is not the best way to measure the input to a dam, but it is an easy way to get an idea of the amount of water being captured. You also need to look into losses from the dams - evaporation, infiltration and use of water (e.g. irrigation) if you want to model the mass balance (e.g. the amount of water captured over a longer periods (e.g. few months)..

You haven't mentioned where your study site is, nor the scale. I was involved in a project in West Bengal that included monitoring of 4 dams in a small catchment (2 km2). We found that the decline in water level through the dry season exceeded the evaporation rate by about 10mm/day - as irrigation use was limited, and the decline was smooth, this was due to infiltration loss (evidenced by the green area downhill of each dam).

Concerning the collection of the data - this could be done using a data logger (maybe even telemetered), or could be using a staff gauge with someone writing down the level each day. Depends on your resources and where the study site is. We trained one of the villagers to collect the daily data, with this supplementing the logged instruments (increasing the coverage of the catchment as well as helping to infill missing data).

Obviously, Barry is right if the dams already exist! I thought you were planning them. Water level is by far the easiest and also direct parameter...

You can use the modified Thiessen polygon method, in which the mean

areal precipitation estimated by the Thiessen polygons is adjusted by using the precipitation gradient calculated by the stations located within or in the vicinity of the watershed

Helo Ammar?

So that your search be very precise, you should know the height of water precipitated per unit area [L / L2] using a standard RWHT. See the attached articles.

Very interesting exchange. It adds to my understanding. I was thinking of ways to trap/retained rainwater in agricultural land. It is very unfortunate we very often limit our concern to the evidence of reducing amount and frequency of precipitation in the tropics and aim at construction of mega-Dam often beyond the retaining capacity of the land.

Thank you all.

I think that the amount of water stored by the dams almost equal the runoff. So, it is possible to estimate the amount of rainfall for each rain gauge starting from the runoff, then go back to other elements of the equation. I mean, you need to calculate evaporation and infiltration, then add them to the runoff to get the total rainfall for each station. Thiessen technique which takes into consideration terrain elevations variations is a very good one for your task.

A SES method of runoff calculation is also available so you can choose this. or for more information u can contact to the department of civil engineering IIT BHU Varanasi professors .

Thanks for all. All answers were interesting and very important...

You may develop a relationship (correlation) between Runoff (R) and precipitation (P) at each location. The correlation is R = aP + b, where a and b are coefficients. Please refer the following website.

http://www.nih.ernet.in/rbis/learning.htm where select "YIELD" under "Runoff" section. This analysis will give a relationship between R and P which can then be used to estimate the runoff at the point interested.

Adham, based on your question I have understood that, you are trying to predict collected amount of rainfall to each RWHT (e.g. reservoir) where there are a number of RWHTs located at each sub-watershed outfall locations covering the entire Drainage Area. If this is the case then you have first understand the basic watershed hydrology that are involved in the Rainfall-Runoff Process. You can not collect total (i.e. 100%) amount of rain water falling over the Drainage Area. There are lots of factors which will control how much amount of rainfall will loss (due to evaporation, initial abstraction - canopy interception, depression storage, infiltration, etc.) and how much rainfall will become direct runoff. Based on the topography, you have to sub-divided entire Drainage Area into a number of Sub-Watersheds. And then you have to place RWHT at each sub-watershed outfall location to capture direct runoff from U/S to D/S area. There are standard hydrologic methodology based on watershed characteristics. Based on your need, you can select one to calculate RUNOFF HYDROGRAPH at each sub-watershed OUTFALL location. The volume of each runoff hydrograph will provide you the volume of rainwater in each RWHT. Now, if you want find out how much percentage of rainfall is received by each RWHT then you can follow the following steps:

(i) Calculate Cumulative Rainfall Depth for the Targeted Rainfall Duration.

(ii) Calculate Rainfall Volume for each sub-watershed individually as follows, Rainfall Volume = Sub-Watershed Area x Cumulative Rainfall Depth

(iii) Calculate % Rainfall collected for each RWHT individually as follows, % Rainfall collected at RWHT = Runoff Hydrograph Volume / Rainfall Volume within sub-watershed x 100