Reliable surveillance of dissolved Rn-222 in arid-region aquifers is challenged by very low natural activity and method-dependent biases, especially humidity sensitivity in electrostatic detectors and air-water partitioning during closed-loop aeration, which can obscure true concentrations needed for defensible drinking-water baselines under preventive frameworks. This study aimed to optimize and field-validate a low-background RAD7 Big-Bottle (RAD H2O) closed-loop protocol tailored for arid conditions and apply it in a regional survey of groundwater used for potable supply in northeastern Saudi Arabia. Groundwater from wells across the region (shallow and deep completions) was collected and analyzed using isotope-resolved alpha spectroscopy (Po-218 and Po-214 windows) with strict chamber humidity control (<= 7% RH), background checks, systematic blanks, duplicates, drift control (+/- 10%), and uncertainty propagation. Air-phase chamber counts were mandatorily converted to water-phase activity using the CAPTURE parameterized by measured loop volumes, temperature, salinity, and humidity, and agreement was evaluated using regression diagnostics and Bland-Altman analysis. The optimized method achieved sub-Bq.L-1 performance, with MDL improving from similar to 0.1645 Bq.L-1 (30 min) to similar to 0.0233 Bq.L-1 (1500 min) and similar to 0.0165 Bq.L-1 (3000 min), and LOQ decreasing from similar to 0.50 to similar to 0.0707 and similar to 0.050 Bq.L-1, respectively. Raw air-phase readings systematically overestimated dissolved radon by similar to 26% (slope approximate to 1.26), a bias removed by the validated air -> water conversion. Surveyed Rn-222 concentrations were uniformly low (0.03-3.20 Bq.L-1), far below commonly used reference values (e.g., similar to 11.1 and similar to 100 Bq.L-1), with no persistent spatial hotspots and broadly overlapping shallow/deep distributions, indicating variability dominated by local lithology and fracture-controlled flow rather than depth. A tiered monitoring scheme is recommended: short screening, routine baselining at similar to 900-1500 min total counting, and similar to 3000 min for ultralow verification, providing a transferable template for sustainable baseline programs in arid aquifers.