Model-of-EQIL-distribution-area

This project aims to create a simulation model for the earthquake-induced landslide (EQIL) area near Zagreb, Croatia to predict the extent (or reach) of the possible landslide (EQIL) triggered by Medvednica mountain’s orogeny.

Landslides are mass movements of soil and rocks along a slope. Landslides occur when the shear strength of the hillside material decreases due to an increase in the shear stress of the slope, or due to processes of change in the natural ecosystem caused by anthropogenic activities (Moresi et al., 2020). Sometimes, landslides can be caused by human activities, or more so, they can modify the threshold of occurrence of landslides by accelerating the dynamics of natural processes (Sidle & Bogaard, 2016). They can cause severe damage, and while most of them occur slowly over time, the most destructive ones happen suddenly after being triggered by an event such as earthquakes or heavy rainfall (Conners, 2019). Earthquake-induced landslides (EQIL) are the most destructive secondary natural hazards associated with earthquakes and are also the focus of this paper (Jibson, 2007).

This research presents a study that aims to estimate earthquake-induced landslide distribution area and to scrutinize its area-wise morphological characteristics. The results show the estimates for the Zagreb/Zagorje region in Croatia. In this region, the Medvednica mountain orogeny triggers seismic events which may potentially lead to earthquake-induced landslides. This research raises the question: What is the EQIL distribution area in the Zagreb/Zagorje (Croatia) neighborhood and what is specific for it in terms of terrain characteristics? To answer it, we estimate the distribution area of an EQIL induced landslide in the region of Zagreb, Croatia, and study its morphological characteristics.

This kind of research is relevant in terms of avoiding the possible effects of such natural disasters. Those effects can include road crashes and debris falling which may cause deaths or destroy properties. Governments and insurers can benefit from estimating earthquake-induced landslide (EQIL) distribution areas by including their risk while developing the infrastructure or selling insurance. Furthermore, residents should also know about the possible dangers of living in such areas.

The structure of the application

The main program uses the functionality of other modules, gradually creating the outputs used in this analysis. The following general overview of the main driver's flow explains the operations performed in order to build the simulation model. Each module has a certain task to perform:

1. mapProximity: create_proximity_map function
   • computes and visualizes the EQIL distribution area for the region of interest.
2. unifyResolutions
   • The resolutions of rasters are unified to follow the smallest, most detailed resolution.
3. cropToDEM
   • The rasters are adjusted to the area of interest (digital elevation model map - DEM)
4. translateToMap
   • The .tif layers are translated to the format recognized by PCRaster library.
   • The outputs are binary rasters on which further operations will be performed
5. extractZone
   • Prepares to include the values only from a certain area, in this case:
     • area within PGA 0.12g contour
     • area of the simulated EQIL distribution area
     • ground failure area estimated by USGS
6. morphologicalStatistics
   • Computes area-wise statistics for the area of interest:
     • Six morphological variables: topographic wetness index (TWI), profile curvature, vector ruggedness measure (VRM), local relief, slope, distance to streams
     • mean, median, min, max, standard deviation, variation are computed
     • The statistical computation contain experimental part performing its operations using Numba just-in-time Python compiler
7. plotMorphology: visualize_csv function
   • produces visualizations of the 6 morphological variables' values distributions. Uses the CSV files containing these values for the 4 areas of interest:
     • area within PGA 0.12g contour
     • area of the simulated EQIL distribution area
     • ground failure area estimated by USGS
     • entire study area

Example results

0	directory	terrain_variable	mean_value	variance_biased	variance_unbiased	standard_deviation_biased	standard_deviation_unbiased	max_value	min_value	median_value
1	PGA_zone	dem	201.94863257257387	17079.648327961608	17080.200477715203	130.6891285760281	130.69124101375425	993.0	82.0	157.0
2	PGA_zone	distance_stream_downstream	13901.147484686871	117626576.87123239	117631116.33468066	10845.578678486105	10845.787953610408	41508.4	0.0	11114.157
3	PGA_zone	local_relief	3.3305738600615915	8.717205243711724	8.717493912009086	2.9524913621739395	2.952540247314012	9.0	1.0	1.0
4	PGA_zone	profile_curvature	0.11128384083136816	3514.897502003044	3515.0129034575716	59.286571008981824	59.28754425220842	6686.0767	-6033.268	-2.7533576e-06
5	PGA_zone	slope	3.5090130031539903	18.6773326575407	18.677938710258513	4.321727971256486	4.321798087631873	30.360462	0.0	1.8577819
6	PGA_zone	TWI_raster	-13.016356720274482	4.404441835135551	4.404584359868914	2.0986762101704852	2.098710165761083	-4.9608107	-17.829576	-13.03805
7	PGA_zone	vector_ruggedness_measure	0.0017229498014403985	1.2117352018472649e-05	1.2117743747435842e-05	0.0034809987099211413	0.0034810549762156646	0.047728363	9.271834e-08	0.00037510128
8	ground_failure_zone	dem	774.4444444444445	3574.135802469136	3784.3790849673205	59.78407649591265	61.51730719860323	876.0	656.0	774.5
9	ground_failure_zone	distance_stream_downstream	15269.94775390625	23684853.1151668	25078079.76900014	4866.7086532036	5007.801889951333	26276.666	12655.301	13266.959
10	ground_failure_zone	local_relief	2.7777777777777777	0.17283950617283958	0.1830065359477125	0.4157397096415491	0.4277926319464987	3.0	2.0	3.0
11	ground_failure_zone	pga_contour_raster_cropped	0.2724381486574809	9.888122256788523e-05	0.00010469776507187848	0.00994390378915068	0.010232192583795444	0.29956156	0.26	0.26934892
12	ground_failure_zone	profile_curvature	6.398951930306238e-07	3.5080181788444543e-13	3.714372189364716e-13	5.922852504363463e-07	6.094564947036594e-07	2.3903008e-06	8.7693635e-08	4.492072e-07
13	ground_failure_zone	slope	24.425234370761448	6.276500191140483	6.645706084736982	2.505294432026001	2.577926702747187	28.02322	19.12525	24.929588
14	ground_failure_zone	TWI_raster	-15.70289765463935	0.4878731211233206	0.5165715400129277	0.6984791486675322	0.7187291144881552	-14.334189	-17.10681	-15.754505
15	ground_failure_zone	vector_ruggedness_measure	0.0065969138457957245	1.113066037581076e-05	1.1785405103799628e-05	0.003336264434335318	0.003432987780898678	0.014293254	0.0027362937	0.0051409993
16	EQIL_distribution_area	dem	720.5886075949367	15239.01430059286	15287.392123769345	123.446402542127	123.64219394595578	981.0	404.0	720.5
17	EQIL_distribution_area	distance_stream_downstream	22621.837090214598	8892256.003877142	8920485.388016436	2981.9885988845	2986.718163472482	27561.291	12633.724	21406.955
18	EQIL_distribution_area	local_relief	4.1835443037974684	3.181501361961223	3.1916013662849094	1.7836763613282605	1.7865053501976729	9.0	2.0	5.0
19	EQIL_distribution_area	pga_contour_raster_cropped	0.2906822777247127	0.0001627926208497987	0.00016330942282075044	0.012759021155629404	0.012779257522280018	0.3195611	0.27278692	0.28980446
20	EQIL_distribution_area	profile_curvature	1.0127504615122613e-05	1.1250372185392921e-08	1.1286087652648136e-08	0.00010606777166223923	0.00010623599979596435	0.001764685	-4.1868538e-05	2.41667e-09
21	EQIL_distribution_area	slope	14.306402396552171	16.316898041568045	16.36869771789048	4.039418032534891	4.045824726540991	25.352772	2.1681957	14.60791
22	EQIL_distribution_area	TWI_raster	-15.105869250961497	2.4253059383153044	2.4330053222464643	1.5573393780147295	1.5598093865105647	-10.349417	-17.52993	-15.439365
23	EQIL_distribution_area	vector_ruggedness_measure	0.01002197717228032	3.7511528433965274e-05	3.7630612651215957e-05	0.0061246655773164686	0.006134379565303728	0.034200232	0.0012128265	0.008357128
24	binary_maps	dem	166.49553249097474	10020.997302544807	10021.14804183545	100.1049314596679	100.10568436325407	993.0	82.0	136.0
25	binary_maps	distance_stream_downstream	15698.401448677438	155459096.3094637	155462729.97461832	12468.323716902112	12468.469431915784	52587.348	0.0	12391.461
26	binary_maps	local_relief	2.462098642833499	6.698877952525168	6.698988826223053	2.5882190696548792	2.5882404884830645	9.0	1.0	1.0
27	binary_maps	pga_contour_raster_cropped	0.13252297116305667	0.004558252954325568	0.004558321521135452	0.0675148350684912	0.06751534285727542	0.33603898	0.042996094	0.1162887
28	binary_maps	profile_curvature	-0.05978386853516508	2318.8650979773934	2318.9022551087724	48.154595813664486	48.15498162297202	6686.0767	-6033.268	-5.759689e-06
29	binary_maps	slope	2.4949044712529873	11.90447348644623	11.904655375756786	3.4502860006738905	3.45031235915776	30.360462	0.0	1.0841638
30	binary_maps	TWI_raster	-12.86445018227667	3.7055675998110598	3.705623464604291	1.9249850908022794	1.9249996011958785	-4.9608107	-17.829576	-12.837674
31	binary_maps	vector_ruggedness_measure	0.0010539718699024707	6.977321516756148e-06	6.977426472027989e-06	0.0026414620036555795	0.0026414818704711923	0.047728363	-2.107556e-08	0.000107259046