I work actually in a research project interested in the estimation of groundwater recharge for alluvial aquifer located in arid area, I need advice on the most accurate methods used in similar areas
It depends of a lot of factors (if you have geological data, boreholes in the zone, budget you have etc) but basically you must start to understand if the rechargea area of this aquifer (for this I do the classical field geological survey + geological map interpretations etc.) and/or if you have boreholes taking water from that aquifer you must do long rate flow test (aquifer pumping test) to see the rate of flow of the aquifer. One of the formula used is Effective infiltraton = rate of flow of the aquifer/ recharge area - Otherwise there are also some new methods using the stable isotopes: you can read different articles among these:
Isotope investigation on groundwater recharge and dynamics in shallow and deep alluvial aquifers of southwest Punjab
and
Choosing appropriate techniques for quantifying groundwater recharge (
You could make use of stable isotopes and chemical methods (chloride, as the most conservative ion) as depth profiles along the unsaturatd zone. For that, you need to collect sediment samples accross the unsaturated zone and extract interstitial water that is trapped and analyse for the aforementionned parameters. Many papers have been published in this regard. Please have a look at the following one, to get an idea on the approaches:
There is a straightforward method based on a simple water balance model and observations from drilling/boreholes. Technical developments are presented in the following paper.
Article Recharge Estimation of Hard Rock Aquifers under Sahelian Cli...
It depends on the availability of data from the study area. Using different rational GIS based methods developed by researchers are among the most easy and cost effective which provides a reliable evaluation data, however application of Stable Isotope and hydrochemical data would be expensive but providing a more reliable information.
Most of time it depends on the availability of data, if you have enough data use one of the hydrological model ( you can read my paper for more details: Coupling WetSpass and MODFLOW for groundwater recharge assessment: case study of the Takelsa multilayer aquifer, northeastern Tunisia ).
There are also GIS- multi-criteria analysis they can give good results