Australian gridded chloride deposition-rate dataset

Abstract

Chloride deposition-rate measurements at points within Australia are upscaled to the entire continent on a regular 0.05° grid. The upscaling uses a double-exponential correlation between deposition rate and distance to the coast, where the parameters in the double-exponential are spatially varying. These parameters are estimated using least-squares with Tikhonov regularisation to ensure minimal spatial variability. A calibration-constrained, null-space Monte-Carlo analysis is used to quantify uncertainty in the prediction. The resulting dataset consists of the best-fit chloride deposition rates across Australia, as well as estimates of uncertainty. The dataset can be used for various purposes including: estimating groundwater recharge through the use of the chloride mass-balance method; catchment salt balance estimates; regional investigations of groundwater hydrochemistry; and, corrosion prediction.

Keywords: Atmospheric chemistry; Chloride mass balance; Groundwater recharge; Hydrochemistry; Rainfall.

Fig. 1

Location and weight of each chloride deposition observation.

Fig. 2

Correlation between the observed chloride deposition and the distance to the coast. The blue and orange lines are fits to the near-coast and central-region data, respectively. The green line is the resulting fit from Eq. (1). The shaded region and $\sigma$ is described below.

Fig. 3

Boxplots of the difference in chloride deposition measured at the same site by difference studies. (Red line is the mean of the absolute difference in the natural log transformed observations).

Fig. 4

Scattered points: the residual, which is the difference between the observation and lnD of Eq. (1). The standard deviation, $\sigma$, of this residual may be fitted with a straight line: $\sigma =0.7+0.00022d$.

Fig. 5

Pilot points used in this study.

Fig. 6

Kriging some artificial data (black dots) using 4 different variograms.

Fig. 7

Results of the uncertainty analysis.