Research Article

Evaluation of the Threshold for an Improved Surface Water Extraction Index Using Optical Remote Sensing Data

Table 4

Summary of the classification accuracy of the AWEI threshold improvement based on the results of the natural break threshold to detect lake surface water, tested at several locations that could represent other lake characteristics in Indonesia.

Test siteThresholdLand cover classUser Accu. (%)Product Accu. (%)KappaOverall Accu. (%)

Maninjau≥−0.232Water97.2499.410.9899.55
Nonwater99.9199.57
Singkarak≥−0.232Water97.1899.740.9899.55
Nonwater99.9699.51
Kerinci≥−0.232Water95.9999.720.9799.43
Nonwater99.9699.38
Rawa Pening≥−0.232Water82.7187.790.7393.58
Nonwater98.4098.01
Batur≥−0.232Water96.4697.750.9799.34
Nonwater99.7199.54
Sentarum≥−0.232Water87.9991.920.7994.33
Nonwater99.6199.39
Jempang≥−0.232Water83.8576.660.7693.83
Nonwater95.5697.15
Tempe≥−0.232Water77.9899.370.8595.77
Nonwater99.8895.14
Matano≥−0.232Water97.6899.180.9899.44
Nonwater99.8299.49
Poso≥−0.232Water98.7799.560.9899.59
Nonwater99.8699.59
Limboto≥−0.232Water91.1693.260.9198.23
Nonwater99.1398.84
Tondano≥−0.232Water94.9798.510.9698.86
Nonwater99.6998.93
Sentani≥−0.232Water97.1992.390.9498.20
Nonwater98.4199.43
Rawa Danau≥−0.232Water----
Nonwater----