Resumen
Design of sampling site water quality monitoring at marine reserve is critical, to optimize the effectiveness of periodic assessment. A simple stratified systematic design, that been usually used in most of monitoring anlaysis, may not maximize the information of spatial data in marine hydrology. The present work applied a multivariate statistical analysis and spatial autocorrelation methods to develop an optimal sampling design for water quality assessment in tropical marine reserve, Banda Sea Conservation Park, Indonesia. Seasonal (west, intermediate, and east monsoon season in Indonesia) and spatial (38 stations) water quality analysis (Salinity, Dissolved Oxygen, pH, and nutrient) in 3 zones of Marine Reserve were conducted. Principal Component Analysis (PCA) showed Dissolved Oxygen (DO) was the principal variable for the sampling design criterion. Spatial DO Variograms showed relocation of the sampling stations, to optimize the design of water quality monitoring. Therefore, even the principal variable may vary at other locations, depends on hydrology and other area specific characteristics, the proposed technique could be applied in sampling design concerning hydrology monitoring.