Hydraulic conductivity and specific yield ( ) are important aquifer parameters, pertinent to groundwater resources management and protection. These parameters are commonly estimated through a traditional cross‐well pumping test. Employing the traditional approach to obtain detailed spatial distributions of the parameters over a large area is generally formidable. For this reason, this study proposes a stochastic method that integrates hydraulic head and time‐lapse gravity based on hydraulic tomography (HT) to efficiently derive the spatial distribution of K and over a large area. This method is demonstrated using several synthetic experiments. Results of these experiments show that the K and fields estimated by joint inversion of the gravity and head data set from sequential injection tests in unconfined aquifers are superior to those from the HT based on head data alone. We attribute this advantage to the mass constraint imposed on HT by gravity measurements. Besides, we find that gravity measurement can detect the change of aquifer's groundwater storage at kilometer scale, as such they can extend HT's effectiveness over greater volumes of the aquifer. Furthermore, we find that the accuracy of the estimated fields is improved as the number of the gravity stations is increased. The gravity station's location, however, has minor effects on the estimates if its effective gravity integration radius covers the well field.

Hydraulic conductivity (K) is crucial for groundwater studies and is conventionally obtained through pumping tests, which are costly due to well installation, resulting in a limited amount of data. Recent studies have addressed this limitation by estimating K through the integration of a pumping-test K and electrical resistivity measurements. While this approach is suitable for local areas, it cannot readily be applied to determine K fields for the composite fan delta. This study proposes and demonstrates an advanced method to estimate the K’s spatial distribution of a composite fan delta. The proposed method included data classification, linear regression, and kriging interpolation. Data classification was conducted using a physical-based zonation method. The K and formation factor (F) data pairs were classified into several groups. Linear regression was used to develop K-F mapping for each group. The regression results showed a good correlation between K and F in each group. These K-F mappings were verified by additional pumping tests. These results indicate that estimation errors were between 7 m/day and 58 m/day, and the correlation coefficient of each data group was greater than 0.8. Based on these regression equations and ordinary kriging method, the detailed K spatial distribution of the study area was derived. According to these results, the proposed method is efficient and economical to delineate K’s spatial features for regional groundwater systems and can benefit groundwater-related studies.