A Comparison on Efficiency of Sampling Techniques for Sediment Estimation in Rivers

Arabkhedri, M., F.S. Lai, I. Noor Akma, M. K. Mohamad Roslan. 2007. A Comparison on Efficiency of Sampling Techniques for Sediment Estimation in Rivers, International conference on mathematical biology, ICMB07, Putra Jaya, Malaysia.

 

Abstract:

In this study, we compared the efficiency of adaptive cluster sampling against SALT (a special kind of probability proportion to size) and simple random sampling for estimating the suspended sediment yield in Gorgan-Rood River, Iran. About 300 sample sets, grouped in different sizes and replicates, were extracted from a five-year daily suspended sediment records, by three study sampling methods including adaptive, SALT and calendar-based. In the next step, total loads were estimated by appropriate sampling estimators; then the loads were compared among themselves and with observed total load as well. The results suggested that all three studied sampling techniques’ average estimates showed satisfactory accuracy. However, adaptive and SALT were more precise than simple random sampling. SALT and adaptive cluster

sampling obtained the optimum accuracy and precision with an average of almost weekly samplings. The result of this study may suggest the application of the adaptive cluster sampling in river monitoring programs.

 

Keywords: Gorgan-Rood River; Suspended sediment load; Adaptive sampling; SALT; Simple random sampling

Estimating suspended sediment load using the adaptive cluster sampling strategy and design approach

Arabkhedri, M., F.S. Lai, I. Noor Akma, M. K. Mohamad Roslan. 2007. Estimating suspended sediment load using the adaptive cluster sampling strategy and design approach. Rivers’ 07- the 2nd international conference on managing rivers in the 21th century: solutions towards sustainable river basins. 6-8 June 2007, Kuching, Sarawak, Malaysia.

 

Abstract: 

River sediment load usually is estimated by using sediment rating curve in combination with continuous flow discharge data of the respective basin. While continuously recorded streamflow data are largely available in gauged basins, suspended sediment concentration records in many monitoring programs are, in contrast, obtained from periodic sampling, usually at predetermined but regular interval time periods over a year or years. It is highly unlikely that this calendar-based sampling or fixed interval approach, carried out due mainly to cost constraints, will ensure the provision of good and representative sediment concentration records because much of the load is carried during a few high flood events. The adaptive sampling strategy considers sampling at more critical river stages, thus increases the level of representativeness of samples. The procedure for selecting sampling units essentially depended on analysis of the sampling survey with consideration and its appropriateness with respect to the population and aggregation tendencies of sediment concentration during transport. To evaluate the efficiency of using the adaptive sampling, a study was conducted for Gorgan-Rood River in Iran with good daily sediment records collected over five years. For comparison purposes, two 93 sample sets (grouped in 31 replicates and three different numbers of records) were selected in accordance to adaptive and calendar-based sampling strategies. The sediment loads for both sampling strategies were subsequently estimated using the sediment rating curve method. Adaptive cluster sampling estimator was used for estimating sediment loads with adaptive sample sets too. Then the estimated loads were compared based on percent error and coefficient of variation. The results suggested that average sediment rating curve estimates with calendar-based sample sets were about 77% lower than the observed output. With adaptive sample sets, the average estimates improved 10-15% better than the calendar-based sample sets. With the adaptive sampling estimator adopted in this study, the sediment load estimation improved markedly, whereby the average underestimation to observed values was reduced to 11%.

Keywords: Gorgan-Rood River; Sediment load; Adaptive cluster sampling; Sediment rating curve