Sampling Designs and Estimation Methods for Sediment Load Prediction in Two Rivers in Iran and Malaysia
The effect of sampling designs and estimation methods on the accuracy of predicted total suspended sediment load (SSL) were evaluated in Gorgan- Rood and Sg. Pangsun, two rivers in Iran and Malaysia respectively. Concerning the rare clustered nature of high load periods during transport, adaptiv...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2009
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/10166/1/FH_2009_10.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The effect of sampling designs and estimation methods on the accuracy of
predicted total suspended sediment load (SSL) were evaluated in Gorgan-
Rood and Sg. Pangsun, two rivers in Iran and Malaysia respectively.
Concerning the rare clustered nature of high load periods during transport,
adaptive cluster sampling (ACS) was conducted. Nevertheless, some
adaptations were made because continuous sediment records of study rivers
are time-scale populations rather than ordinary spatial-scale in previous
studies. This study suggests the use of forward neighborhood relation
instead of symmetric and the use of flow duration curve instead of ranking of
initial samples.
In order to evaluate the ACS in river sediment estimation, 1 2000 sample sets
were generated in each river which categorized in scenarios differed in terms
of neighborhood relation, flow discharge threshold, cluster size and initial
sample size. For all sample sets, total SSLs were predicted using the modified Hansen-Hurwitz (HH) and Horvitz-Thompson (HT) estimators which
then were statistically evaluated. The results suggests that ACS can not
estimate sediment load properly when sample size is smaller than 15 % of
the size of respective population. HT estimator showed a better performance
than the HH. Moreover, employing forward neighborhood relation instead of
symmetric showed underestimations less than 5%. |
|---|