2,798
Views
7
CrossRef citations to date
0
Altmetric
Articles

Forecasting landslides using SIGMA model: a case study from Idukki, India

, , , , , & show all
Pages 540-559 | Received 16 Jul 2020, Accepted 28 Jan 2021, Published online: 14 Feb 2021

Figures & data

Figure 1. Location detains of Idukki (a) India (b) Geological map of Idukki district (modified after (Geological Survey of India Citation2010)).

Figure 1. Location detains of Idukki (a) India (b) Geological map of Idukki district (modified after (Geological Survey of India Citation2010)).

Figure 2. Location of rain gauges, landslides and reference areas considered for the analysis.

Figure 2. Location of rain gauges, landslides and reference areas considered for the analysis.

Figure 3. Annual and cumulative rainfall recorded in the four rain gauges during the study period.

Figure 3. Annual and cumulative rainfall recorded in the four rain gauges during the study period.

Figure 4. Data transformation using target function for Idukki.

Figure 4. Data transformation using target function for Idukki.

Figure 5. An example of threshold curves for cumulative rainfall up to 100 days; ‘σ’ is the standard deviation of each series.

Figure 5. An example of threshold curves for cumulative rainfall up to 100 days; ‘σ’ is the standard deviation of each series.

Figure 6. Decisional algorithm used for calibration.

Figure 6. Decisional algorithm used for calibration.

Figure 7. A graphical representation for the process of optimization for increasing the threshold value.

Figure 7. A graphical representation for the process of optimization for increasing the threshold value.

Table 1. Optimized threshold and N values for each reference area.

Figure 8. Number of false alarms before and after optimization.

Figure 8. Number of false alarms before and after optimization.

Table 2. Results of validation of SIGMA model for Idukki.

Figure 9. Correctly predicted landslides (TP) and missed landslides (FN) during validation.

Figure 9. Correctly predicted landslides (TP) and missed landslides (FN) during validation.