Original paper

Landslides (2021) 18:1597–1610 Mysuru R. Yadupathi Putty I B. N. Prithviraj I P. Nalina Kumar I M. G. Nithish I
DOI 10.1007/s10346-020-01589-y Ghanshyam Giri I P. N. Chandramouli
Received: 17 June 2020
Accepted: 18 November 2020
Published online: 7 January 2021 An insight into the hydrological aspects of landslides
© Springer-Verlag GmbH Germany
part of Springer Nature 2021 of 2018 in Kodagu, South India

Abstract The goal of the work has been to assess the factors that an unprecedented number and sizes of landslides devastating the
led to the series of landslides, which ravaged the District of southern Western Ghat regions of the state of Karnataka. Indian
Kodagu situated in the mountainous Western Ghats, in 2018. The Space Research Organisation (ISRO 2018) identified over 900
work has been accomplished through field investigations and slides and furnished the information to the Geological Survey of
analysis of field data. While commonly, studies in India lay em- India (GSI) and other disaster management agencies. Many of
phasis on the geological aspects and anthropogenic factors asso- these slides occurred during the second week of August, but most
ciated with landslides, the present work intended to study in detail of the slides occurred on the 16th, following 3 days of extremely
the hydrological aspects of the disaster, which wreaked havoc in heavy rainfall. GSI (2018a) documented 105 instances of significant
the region characterised by deep soils, very heavy rainfall and a magnitude from one part of the small district of Kodagu, which
dense vegetation cover. All the slides of significant magnitude spans 4108 km2 well inside the mountains. Many more cases of
along a 25-km stretch of the Payaswini valley, draining the western slides of nature similar to those surveyed by the GSI have been
slopes of the Madikeri hill range, were identified, and their land reported from other parts of Kodagu. The landslides in Kodagu
characteristics delineated using field surveys and maps. Soil char- include many, probably over twenty, that comprised removal of
acteristics of the region were determined through tests on samples quantities of earth exceeding 25,000 m3, which amounts to the
collected from ten slide sites. Results from tests on texture, density movement of soil mantles measuring (Parkash 2019) like 200 m
and permeability were used to learn about the part played by long × 50 m wide × 5 m deep. Some of the slides involved earth
subsurface water. Characteristics of rainfall during the year were measuring over 100,000 m3 also. Obviously, these natural disasters
investigated using field data from the worst affected region. Stud- are a subject of great research interest. The GSI carried out pre-
ies on progress of the Monsoon rain, an analysis of probability of liminary studies on those which were approachable, located most-
extreme values and establishment of intensity–duration relation- ly on the highways. It has come out with a report (GSI 2018a),
ships were accomplished. A variable source area watershed model sighting human interference as the most important factor leading
incorporating MODFLOW, the popular ground water flow simula- to the disaster (SOM 2018b). Later, GSI (2018b) has come out with
tion model, has been applied on a typical micro-catchment in one more report, analysing a few selected landslides along two
order to study the growth of the saturated zone during rains. It highways in the district. The GSI reports also furnish measures to
has been inferred from the investigations that an unprecedented reduce the possibility of more of landslides in future. But, these
combination of extremely wet catchment and a very heavy spell of studies are basically Geology oriented and generic ones suggesting
rainfall, rather than any commonly attributed factor, has been the well-known measures that should be taken up during unavoidable
primary cause of the disaster in the district. It is concluded that it human interventions in nature. Further, Jain et al. (2019) have
would be a futile exercise working on this natural phenomenon attempted to analyse the disaster and map the potential hazard
devoid of extensive field studies, because of the uniqueness of the zone using remotely sensed data and GIS approaches. But, this
region. exercise is based on GSI reports and on general perceptions re-
garding the causative factors and contain none on the site inves-
Keywords Deep soil landslides . Factors influencing tigations. Other studies (e.g., IEI, Mysore Centre, as reported in
landslides . Field investigations . Model of saturated source SOM 2018a; BRO 2018) reported in the aftermath of the disaster
area . Uniqueness of Western Ghats have focused on the extent of the damage caused to property and
civic amenities, with a focus on steps to be initiated to set things
Introduction right.
The Western Ghats, locally called the Sahyadri ranges, are a series Research studies on landslides in Western Ghats reported in
of hill ranges bordering the west coast of India, almost throughout literature are few and far apart. A major part of the literature
its length. They run about 1600 km nearly parallel to the coast, available focuses on application of Remote Sensing and GIS tech-
with an extremely narrow east-west profile (Fig. 1). The Sahyadris nologies for disaster analysis and hazard zone mapping (Kanungo
consist of mountain ranges with an average elevation of 1000 m, et al. 2009; Scaioni et al. 2014). Such studies by Kumar (2005),
displaying an extremely steep western face and gently descending Mishra et al. (2010), Ramachandra et al. (2012) and Maruthi and
eastern slopes. The region lies in the tropical South Asian mon- Chandran (2013) pertain to Kodagu or nearby areas in the
soon tract, characterised by extremely wet summers and dry win- Sahyadris. Field studies on landslides in this region have been
ters. Almost through the length, the annual rainfall in the areas reported by Ramachandra et al. (2009), Sajinkumar et al. (2011)
near the crest of the ranges exceeds 5000 mm, with many locations and Senthilkumar et al. (2017). While the latter two studies focus
receiving over 7000 mm. Over 90% of this occurs in the south-west on the soil and geotechnical aspects of the slides, the former
monsoon period spread over the months June to September. The focuses on the anthropogenic activities leading to the disaster.
combination of steep slopes and very heavy rainfall renders the However, no detailed studies on the hydrological aspects, includ-
region susceptible to landslides. The monsoon months of 2018 saw ing rainfall and subsurface flow of water, which are among the
Landslides 18 & (2021) 1597
 Original paper
pressure development leading to structural instability; there
have also been a few studies on field using simulated rainfall
(e.g., Moriwaki et al. 2004). Experiments on piping (Zhenlei
et al. 2019) in this regard are particularly notable.
(ii) Use of hydrological models for analysing subsurface flow and
ground water build-up (e.g., Li et al. 2015; Reder et al. 2017;
Abbate et al. 2019);
(iii) Instrumenting and monitoring landslide prone sites to un-
derstand the mechanism and characteristics of the slides
(e.g., Huntley et al. 2019; Maneesha and Vasudevan 2012);
(iv) Landslide modelling and development of Local Landslide
Early Warning Systems, based mostly on thresholds for rain-
fall intensity and measured soil moisture or antecedent rain-
fall quantity. It may be noted that the below-mentioned
thresholds, rainfall beyond which would lead to landslides,
have been suggested for different regions by various investi-
gators: five daily rainfall of 320 mm in Vietnam (Chau et al.
2019); 5-day rainfall of 189 mm and a 2-day rainfall of 98 mm
in China (Zhao et al. 2020); 5-day rainfall of 350 mm in
Uttarakhand Himalayas in India (Martha et al. 2014) and a
3-day value of 450 mm in Sri Lanka (Khang et al. 2019).
Interestingly, the 5-day rainfall of 350 mm is so common in
the areas being dealt with in the present study that its
recurrence interval is less than 2 years and even the 3-day
value of 450 mm occurs once in less than 5 years in some
parts of the region.>

Despite the above advances, most of the studies in India aim at

analysing data on certain predecided factors supposed to influence
or trigger landslides (Anbalagan 1992) and emphasise on anthro-
pogenic factors while suggesting mitigation measures
(Ramachandra et al. 2012; GSI 2018b; SOM 2018a). Such studies
overwhelmingly assume a few preconceived ideas about the mech-
anism of landslides and about hydrological processes to hold good
uniformly everywhere. For example, it is normally assumed that
the short-duration rainfall intensities are a direct function of the
monthly and annual values, that flood flow in streams is all due to
overland flows and that subsurface water contributes only to the
base flow. Also, for convenience, these studies use readily available
maps on lineaments, soils, land-use and land-cover (LULC), rain-
Fig. 1 Western Ghat ranges bordering the west coast of India
fall etc., which usually are on a scale not compatible with the scale
of the images. As a result, predictions are done beyond the fun-
damental understanding of the processes (Burt and McDonnell
2015). But, chances are always there that the generalised and
primary factors leading to deep landslides (Van Asch et al. 1999) preconceived ideas regarding the various natural processes may
have been reported from this region. The present work is an not hold correct in every case (Jackson 1988). Further, there have
attempt to fill in this gap, accomplished through field investiga- been contradicting views too in generalising theories. For example,
tions. The work has been carried out by analysing data on slides in while Sharma et al. (2012) consider soils deeper than 1 m as stable,
one particular stretch of land in Kodagu and on characteristics of Balasubramanian (in The Hindu 2018) terms areas with soils
soil and rainfall in the intensely affected area (Fig. 2). deeper than 1.5 m a zone of uncertainty. The Western Ghats are a
classic example of the ‘exceptional’ cases, where the runoff mech-
anisms are totally different (Putty 2006) from those characterising
Research Trends and the Present Work the mountainous areas elsewhere in the country. Hence, concerted
Recent research on landslides seems to have largely been oriented studies, including field investigations as an indispensable part,
towards the following: need to be undertaken on landslides in this region, in order to
understand their mechanism. The present work was initiated in
(i) Experimental studies (e.g., Giri et al. 2018) on the mechanism this background, focussing on the hydrological aspects of the
of slides with emphasis on subsurface flow of water and pore- landslides of 2018 and using field data procured from the sites

1598 Landslides 18 & (2021)

Fig. 2 Location of the study area in the district of Kodagu, Karnataka

for analysis. The aim was to understand the characteristics of the the west flowing Barapole and the east flowing
slides and to establish the factors that led to the deluge during the Lakshmanatirtha and Kaveri;
week spanning 11th to 17th August 2018. The investigations com- (ii) The Pushpagiri range in north running nearly south-north,
prised the following: forming the divide line between Kumaradhara and the east
flowing Harangi and Hemavathi;
(1) A reconnaissance survey of all the slides adjoining the Na- (iii) The Benganadu (Madikeri-Galibeedu) range running nearly
tional Highway NH-275 and the slopes draining into the valley normal to (i), interlinking the above two, and forming the
of the river Payaswini (in which NH-275 runs) in between ridge line between the west flowing streams Payaswini and
Madikeri and Sampaje, a distance of about 20 km, in order to Kumaradhara and the Kaveri;
establish the most commonly affected landscape; (iv) A few other minor ranges, branching off from the main
(2) An intensive study of the selected sites to understand the ranges.
topographic features and other characteristics that influence
the vulnerability of land;
(3) Studies on selected ten landslide sites to understand the soil As is the case with the whole of the central and southern
characteristics in the region; Western Ghats, Kodagu is underlain by Precambrian Igneous
(4) Studies on the rainfall records pertaining to 2018 in order to rocks, intrusive granites and granitic gneiss forming the basic
establish the severity of rainfall that led to the disasters; and, types (Pascal 1984). The district experiences annual rainfall in the
(5) A study on the variations in the location of the ground water range of 1700 to 7200 mm, with an average of about 120 rainy days/
table and the growth of the saturation zone in a typical year all over. Over 80% of rain in the district is of the orographic
catchment during rainfall, using a watershed model.> type (Putty 2010) resulting due to a continuous north-east flow of
moisture from the Indian Ocean and the Arabian Sea obstructed
by the mountain barriers. As a consequence, even though the daily
The details concerning the methodology adopted and the work rainfall depths are very high, short-duration intensities are only
done under each of these cases is furnished after presenting a moderate.
description of the study area, the district of Kodagu. The rare combination of very heavy moderate intensity rainfall
and a hard rock base material has resulted in development of a
The study area huge regolith of soil (Bourgeon 1989) in the region—the mantles
Situated in the cradle of the Western Ghats, Kodagu is ‘a land of are often greater than 50 m in depth in mid-slope areas (Darshan
million hills’ (Richter 1870). The district is characterised by land 2016). Also, the streamflow in the region is dominated by subsur-
with elevation varying between 900 and 1750 m (Kodagu District face runoff. The mechanism of runoff generation here is best
Disaster Management Plan 2017–2018).The following mountain explained by the variable source area theory (Putty 2010), accord-
ranges run through the district (Fig. 3): ing to which overland flow is limited to saturated areas riparian to
the streams and to small patches of land devoid of vegetation.
(i) The Brahmagiri range running south-east–north-west all The district, a major portion of which probably was covered
along the Kerala border and forming the ridge line between with rainforests prior to the nineteenth century (Richter 1870) has

Landslides 18 & (2021) 1599

 Original paper

Fig. 3 The major mountain ranges and drainage basins in Kodagu

undergone a lot of land-cover changes. There have been appre- & Topographic feature of the site—valley, with contours concave
hensions of these changes casting ill effects on the hydrology and d/s; the watershed divide, with contours convex d/s; or the
the ecology of the land since the period of the British (Bidie 1869). mid-slope, with almost straight contours;
An unaccounted extent of prime forest land seems to have been & Length and area covered on the image by the slide portion and
exploited for commercial crops, coffee in particular. Still, the the area affected d/s;
district is host to a dense vegetation cover. & Distance of the highest point (the crown) from the road, if
All the above features, together with steep slopes, render the found nearby;
area landslide prone. Yet, landslides have not been a very regular & Distance of the crown of the slide from the ridge;
feature, except for the road-side mud falls that disrupt traffic for & Gradient upslope (from the crown to the peak/ridge)—computed
just a few hours, in the district. The slides of 2018 have been as the ratio of the difference in RL (called relief) to the distance
exceptional, both in number and sizes. Hence, a study on them between the two points;
is interesting. The area most affected during the year lies mainly in & Local slope obtained using the DEM;
the Benganadu hill range, covering mainly the sub-basin of & Land cover and human interference at the crown and in a
Harangi and upland areas of the basin of Payaswini (Fig. 2). major portion of the slide. This information was obtained from
a study of the topo-sheets, the Google Earth and a reconnais-
Locational features of the sites affected sance survey of the region approachable. The land cover is
In order to understand the features of the land that has been most classified as forest, scrub, tree clad area, plantation and barren/
commonly affected, location of all the landslides sites adjoining urban. The density of vegetation is also included in terms the
NH-275 and the Payaswini valley was delineated and investigated words dense, open and sparse. Human interference is identi-
(Fig. 4). Only those slides which could be identified on ‘Google fied in terms.
Earth’ were included in the study. A reconnaissance survey was
also carried out in order to ascertain the location of these sites. It
was observed during the survey that many small slides, mostly in Delineation of these features was accomplished using the Goo-
the valleys and a few others adjoining the road, could not be gle Earth, the Survey of India Topo-sheets and a 30.5-m resolution
recognised on the image. These were not included in the study. DEM furnished by NRSC, ISRO. A total of 25 landslides were
Those in the valleys have gone out of the reckoning since they identified and listed. The above features pertaining to the sites
remain covered by thick vegetation that characterises the valleys are presented in Table 1. Figure 4 shows the location of these sites.
and are away from the roads. The following features pertaining to The following are the important inferences that can be drawn from
the so spotted landslides were delineated: the illustrations:

1600 Landslides 18 & (2021)

Fig. 4 Location of the sites investigated

& The general perception that landslides are caused by anthro- & As mentioned earlier, many more slides that occurred in the
pogenic activities like land-cutting, for road building in par- valleys during the critical period of 2018 could be sited during
ticular, is ill founded—most of the landslides studied are the reconnaissance. This fact highlights the importance of
found to have been initiated far removed from the roads. Also, subsurface water in causing slides. The present work gains
those adjoining the road are very small in size. importance since it focuses on hydrology pertaining to the
& Land slope at all the sites, except a couple, is very high ranging deluge of 2018. Rainfall, topographical characteristics and sub-
from about 30% to over 60%—the study area is on the western surface flow mechanisms are the aspects of hydrology that can
slopes of the ranges, which in general is very steep, and hence be supposed to influence the vulnerability of the land to slide.
no inference could be drawn specifically on the influence of Field surveys show that all the landslides have been removal
slope on the slides. and movement of huge masses of soil, along with the vegeta-
& Sites with all sorts of land cover, including grasslands, open tive cover on the surface and that the debris consisted of very
scrub, plantations and dense forest lands, have been affected. few rock masses. Hence, investigations on soils carried out in
This fact opens up a Pandora’s Box—it renders the argument selected sites will be presented first. Rainfall and other hydro-
that vegetative cover with deep-rooted trees and shrubs can logical aspects of the deluge will be looked in to subsequently.
reduce the occurrence of rapidly moving landslides by
strengthening soil layers and improving drainage. There have
been arguments (Hamilton 2008) that Deep landslides
resulting from continuous heavy rainfall or earthquakes are Soil characteristics
less likely to be affected by vegetation. According to Bruijnzeel A glimpse of the depth of soil in the region can be had from the
(2004), whereas a well-developed tree cover can prevent shal- road side cuttings on highways and from the vertical cuts caused
low landslides up to about 1 m in depth, landslides of depth due to the landslides—which are often deeper than 10 m. The
more than 3 m are not appreciably affected by forest cover. The information about soil mantles in Kodagu made available by the
present one is a very strong case in support of these arguments. Dept. of Mines and Geology is presented in Fig. 5. The subsurface
It is interesting to note that vegetation cover increases infiltra- profiles shown here are from places namely Madapura,
tion, which only helps build up of the ground water table in the Somavarapete and Napoklu. Places around Madapura experienced
valley portions, enhancing the chances of sliding. Facts also go innumerable slides in 2018. It is seen that soil thickness at these
to support the theory that hefty trees on unsupported cut three places is as high as 22 m, 28 m and 42 m, respectively, and the
slopes even may go counterproductive (Forbes and hard base rock is present only below these depths. Soils so deep
Broadhead 2011), often leading to small land slips. In cases of are not exclusive to Kodagu alone. Data on subsurface profiles
extreme climatic events such as hurricanes, trees may actually furnished by the Karnataka Power Corporation Ltd., obtained
increase slope instability due to the trees’ weight and the from their project sites, confirm that deep soils feature all highly
susceptibility of particularly high trees to uprooting due to wet regions of Western Ghats in Karnataka (Darshan 2016). Sankar
extreme storms, damaging the soil matrix (Bruijnzeel 2004). et al. (2017) have detected and investigated tunnels produced due
& Very few landslides are located near the watershed divide and to soil piping 30 to 40 m below the ground level in the district of
Kannur, in the state of Kerala (Fig. 1), adjoining Kodagu. This
ridges; valleys are the zones prone to slides, meaning that
hydrology is the subject matter most pertinent to a study of explains why landslides in the region have been removal of only
landslides in this region. This observation is in conformity soil. Sajinkumar et al. (2011), who have carried out preliminary
with the experiences reported by Kuriakose et al. (2008), who studies on landslide occurrences in Ernakulam and Idukki dis-
have carried out extensive studies on landslides of Kerala. tricts of Kerala, also opine that presence of very thick weathered

Landslides 18 & (2021) 1601

 Original paper
Table 1 Locational, topographic and land cover characteristics of the sites
No. Topography Lengtha Areab Dist.c Landd Human Dist.e Slopef Local
(m) (ha) road cover interference ridge (%) slope
(m) (m) (%)
1 Mid-slope 13 6 12 OS D 29 44.8 55.8
2 Valley 20 4 166 OS D 103 34.0 49.0
3 Mid-slope 123 5 117 TC-S D 31 26.9 56.6
4 WS divide 100 14 155 BU D 71 35.2 40.4
5 WS divide 29 12 30 TC-S D 271 28.5 51.0
6 Valley 67 6 195 TC-S D 307 29.6 28.7
7 Mid-slope 18 4 28 TC-S D 495 29.7 30.6
8 Valley 910 444 1016 OS+ Pl. D 248 21.4 32.5
9 Valley 738 202 1059 OS PD 230 39.1 21.3
10 Valley 31 28 39 Pl. D 129 30.2 61.3
11 Valley 943 974 800 TC-D PD 64 34.4 55.4
12 Valley 447 179 454 Pl. D 330 36.4 22.2
13 Valley 1181 1143 4042 TC-D SD 990 25.5 30.6
14 WS divide 2040 751 1852 OF/ DS SD 53 15.1 21.3
15 Mid-slope 1885 404 1862 DS U 40 17.5 34.4
16 Mid-slope 140 61 347 Pl. D 1550 26.9 44.5
17 WS divide 247 90 584 Pl. D 1490 32.9 48.3
18 Valley 2276 1044 2314 DMJ U 685 36.6 36.0
19 Mid-slope 78 20 42 TC-D PD 160 45.0 40.4
20 Valley 206 42 340 TC-D PD 2246 24.8 44.5
21 Valley 798 394 1230 DMJ U 1516 31.9 40.8
22 Valley 2180 1761 2290 DMJ U 540 53.3 48.1
23 WS divide 2743 1175 2594 DMJ U 810 49.9 73.2
24 Valley 1846 813 2094 DMJ U 56 38.8 60.1
25 Valley 183 75 2360 DMJ U 1278 32.0 14.1
a
Length of the landslide plus the area affected, as seen on the image
b
Area affected in square meters
c
Distance of the highest point (crown) from the nearest road in meters
d
Land cover in the vicinity of the slide
e
Dist. of the crown from the ridge (m)
f
Average slope from nearest peak/ridge to the crown
Notations: FDS, fairly dense scrub; OS, open scrub; DS, dense scrub; Pl., cardamom/coffee plantation; TC-S, sparsely tree clad; TC-D, densely tree clad; DMJ, dense mixed jungle
(forest); B/U, barren/urban; D, disturbed (regularly accessed); PD, partly disturbed (often accessed, but not much activity); SD, slightly disturbed (seldom accessed); U, undisturbed
(no access)

zones is the main cause of landslides in the region. Detailed Commonly, studies on deep slides presume that landslides are a
studies on landslides in such deep soil regions are not available direct result of structural deficiencies in the strength of the rock
and general studies on landslides attach only a secondary impor- and emphasise on the presence of weak zones like faults and
tance to soil characteristics (Sharma et al. 2012; Fonseca et al. consider lineaments an important factor in mapping hazard zones
2017). Most studies deal with soils of thickness of the order 1 to 2 (Anbalagan 1992; Chawla et al. 2018). Therefore, a study on land-
m and classify them as shallow (< 50 cm), moderately shallow, slides in Kodagu is important—it can be expected to throw new
moderately deep and deep (> 100 cm). Generally, soils on granite light on this natural disaster in very deep soils.
are believed to be shallow with low storage and high permeability, In the present work, soil characteristics of the region of the
resulting in relatively high slope instability (Walker and Shiels landslides were analysed using soil samples collected from ten
2013). Hence, deep landslides are considered rooted in bedrock. slide sites. These sites were selected through a reconnaissance

1602 Landslides 18 & (2021)

 of soil water storage during prolonged rainfall spells, leading to
excessive pore pressures.

Extensive studies in the region (Putty and Prasad 2000; Putty

2006; Sankar et al. 2017) have established that the soil mantles in
the Sahyadris are characterised by pipes that drain subsurface
water quickly to the streams. Piping renders the slopes stable
(Jones 1994) in the usual case. But, under exceptionally heavy
continuous spells of rain, the soil matrix may get saturated in
the valley portions, into which soil moisture converges. An inves-
tigation of the rainfall of 2018 and an analysis of the soil water
storage pattern on a typical slope can hence be expected to reveal
the cause of the devastation that struck Kodagu during the year. A
discussion of the rainfall pattern is presented first.

Studies on the rainfall of 2018

The work on rainfall included (i) studies on progress of monsoon
rainfall in the year, with data on cumulative rainfall, (ii) probabil-
ity analysis of indices of extreme rainfall and (iii) analysis of short-
Fig. 5 Subsurface profiles showing the soil depth and type in the study area duration intensities on the days of the deluge. The first two
(Source: Dept. of Mines and Geology, Madikeri) components were accomplished using daily rainfall data, made
available by Karnataka Directorate of Economics and Statistics
(KDES), from stations in the affected area. Intensity studies
survey in the region keeping in view the facilities available for pertained to the data from the Self Recording Rain Gauge
collection and transport of the samples. At each of the ten selected (SRRG) at Kudigana being maintained by NIE-WRC.
sites, soil samples were collected at two points—one near the
crown and the other near the base. At the crown, samples were
collected on the surface. At the base, sampling was done at undis- Progress of the monsoon rain
turbed portions of the land nearest possible to the slide, so as to It has been observed that the catchment of Harangi (Fig. 2) was the
get information on the in situ soil. Samples were collected 3–6 m most affected due to the excess rainfall of 2018. Hence, studies on
below the surface either on the vertical face of the slide or at tall rainfall pertained to the data from this region. Daily data from
road cuttings adjoining the slide. Both core and bulk samples were stations at Surlabbi, Galibeedu and Kudigana were used. It has
collected to facilitate laboratory tests for density, permeability and been found from the records that since 2000, 2013 and 2018 have
texture. been the years of excess rainfall while, 2011 has been the one
Results from the tests are presented in Table 2. This table recording the normal amount. Hence, the rainfall series during
furnishes the textural classification, the fraction of silt and clay these years were mutually compared by plotting cumulative
present, the density and the permeability of the soil. A few impor- rainfall against time. Figure 6 shows the facts. It is clearly
tant observations from the table are as below: observed from the illustration that the progress of rainfall in
2013, which received almost the same amount as in 2018, has
& Soils are dominated by the silt fraction—soils at the bottom been very steady, whereas 2018 saw an extremely heavy spell
during the period between the 5th and 16th of August, which led
layers are all sandy silts, while those at the surface are either
sandy silts or silty sands; to the greatest damage. This inference is in coherence with the
& Clay percentages are low—they vary between 2.3 and 12.7 in the observations of Ibsen and Casagl (2004) that a two-stage pattern of
precipitation, comprising a preparatory period and a triggering
bottom layers. This renders the soil highly stable, permitting
vertical unsupported road cuttings; period, is necessary for inducing most landslides.
& Dry densities are generally low, ranging between 900 and 1400
Probability analysis of extreme values
kg/m3. Considering the specific gravity of solids as 2.6, the
corresponding porosities work out to be between 65 and 45%. Most of the landslides in the region were reported to have oc-
This means the soils are highly porous, capable of storing large curred on the 15th, 16th and 17th of August. Hence, an attempt was
quantities of water; made to quantify the rainfall of this spell by means of a probability
& Soils are generally highly permeable at the surface, the values analysis, using records from four stations in the Harangi catch-
ment and two other prominent stations in Kodagu. For the pur-
of coefficient of permeability (K) going up to 250 mm/h.
Results from infiltration tests reported by Putty et al. (1997) pose, indices of extreme rainfall, 1-day, 2-day, 3-day, 5-day and 10-
and Chaitra et al. (2015) support this observation. However, day extreme values were computed for sufficient lengths of period
permeability values in the subsurface layers indicate that (Table 3), and the series were subjected to probability analysis.
drainage rates are low, ranging between 1 and 4.8 cm/h. The Suitability of the normal, log-normal and Gumbel distributions
high infiltration capacity of the terrain along with a low per- was studied on data from the station of Somavarpete, whose
meability soil matrix can be expected to result in high amounts record length is over 100 years. It was found that the Gumbel

Landslides 18 & (2021) 1603

 Original paper
Table 2 Soil characteristics at landslide sites
Sampling site Silt (%) Clay (%) Texture Density (g/cm3) Ks (cm/h)
Site 1
Top 44 9 Sandy silt 1.35 4.6
Bottom 36.7 6.5 Sandy silt 1.06 4.8
Site 2
Top 27 7 Gravelly sand 1.35 19.6
Bottom 36 3 Sandy silt 1.28 2.1
Site 3
Top 36.6 10.2 Sandy silt 1.13 8.6
Bottom 39.6 2.4 Sandy silt 1.37 1.1
Site 4
Top 29.6 14 Sandy silt 1.00 25.3
Bottom 21.8 9 Sandy silt 1.03 2.3
Site 5
Top 51 11 Silty sand 1.24 14.7
Bottom 57 12.7 Silty sand 1.25 3.5
Site 6
Top 27.7 3.5 Sandy silt 1.27 13.3
Bottom 32.8 2 Sandy silt 1.37 1.6
Site 7
Top 27 6 Sandy silt 0.9 4.3
Bottom 31.5 10.5 Sandy silt 1.3 1.3
Site 8
Top 32.2 10 Sandy silt 1.05 20
Bottom 32.4 8 Sandy silt 1.39 2.1
Site 9
Top 27.2 5 Gravelly sand 1.2 5.4
Bottom 38.7 9.5 Sandy silt 1.2 4.6
Site 10
Top 33.1 11.9 Sandy silt 1.27 13.3
Bottom 30.5 2.3 Sandy silt 1.37 1.6

distribution fits the data the best (Nitish 2019). Hence, the return Interestingly, few slides have been reported from this region. Further,
periods of extreme rainfall for the year 2018 were estimated using data from the station at Kudigana shows that the 5-day extreme for
the Gumbel function. The results are presented in Table 3. the year (during the period 13th to 17th of August) is equal to
It is observed from the table that for the stations in the Harangi 1686 mm and is the highest ever recorded in the history of Kodagu.
basin, while the 1-day, 2-day and 3-day extreme values for 2018 show The earlier highest 5-day extreme for any location in the District has
return periods of 3 to 28 years, the 5-day and 10-day extremes have been 1462 mm, recorded in 1964 at Bhagamandala, a station receiving
been highly excessive with return periods exceeding 50 years. It is rainfall similar to Kudigana. This spell of unprecedented rain fol-
also seen that in the case of Galibeedu, the 2018 rainfall has been lowing a highly wet 2-month period seems to have been behind the
exceptionally high, with the extreme values showing return periods devastation caused in the region. Yet, the short-duration intensities,
of 150 to over 500 years. Consequently, it has been observed that it is which are always higher than the intensities for longer durations, are
the region close to Galibeedu (Fig. 2) which was the worst affected by usually considered to be more important in causing overland flows
landslides. On the other hand, data from Bhagamandala, which is and landslides. The following is an analysis of intensities on the 3
situated in the wettest part of Kodagu, shows that the 2018 rainfall in days of the deluge. The data analysed has been procured from the
that region is not exceptionally different from the normal year. SRRG being maintained by the NIE-WRC at Kudigana.

1604 Landslides 18 & (2021)

 7000 rainfall intensity is higher than the infiltration rates. The
Cumulative RF (mm) 6000 duration for which the various intensities were exceeded on
5000 Galibeedu the most intense rainy day is shown in Fig. 9. It can be
2011 observed from here that intensities greater than 50 mm/h
4000
2013
3000 2018 lasted only about 10 intervals (of 15 min) on the day. Con-
2000 sidering the infiltration rates of the order of 100 mm/h, as
1000 explained in ‘Soil characteristics’ above, it can be seen that
0
most of the rainfall gets infiltrated in this region. Hence,
1 11 21 31 41 51 61 71 81 91 excessive overland flow cannot be expected to have been
Day number the cause of the landslides.

9000
Cumulative RF (mm)

8000 On the other hand, high rates of infiltration lead to excessive

7000 Kudigana storage and result in flow of large quantities of subsurface water
6000 2011 on highly wet days. Subsurface flow from the landslide locations
5000 2013
4000
has been documented by GSI (2018a) also. As reported by Jain
2018
3000 et al. (2019), GSI states thus: ‘Clean water flow in force from inside
2000 the slope: it was observed that considerable volume of very clean
1000 water was coming out from the middle parts of the landslides. This
0
1 11 21 31 41 51 61 71 81 91
continuous flow of the water in force inside the slope in fissures
and cracks will decrease the stability of the slope and eventually
Day number will act as a triggering factor for slope failure’. Probably, GSI has
wrongly attributed ‘continuous flow of the water in force’ to flow
8000
through fissures in rocks. Studies (Jones 1994; Tsukamoto et al.
Cumulative RF (mm)

7000
1982) have established that piping, in spite of being a mechanism
6000 Suralabbi
of draining water out quickly, results in the soil mantle near the
5000 2011
2013 valleys grow weaker than the rest of the land. Under prolonged
4000
2018 heavy downpour lasting days together, as it happened during the
3000
second week of August 2018, the valleys may get saturated and
2000
yield under extreme conditions. The simulations of a hydrological
1000
model can be used to throw light on the growth saturation in
0
1 11 21 31 41 51 61 71 81 91 valleys. An attempt in this regard is presented below.
Day number
A model to study subsurface flow and surface saturation
In order to investigate the build-up of the water table and the
Fig. 6 Progress of monsoon rainfall (from 1st June to August end) of 2011
(normal), 2013 (highly wet) and 2018 (highly wet) in the Harangi basin growth of the saturated zone during rains, the simulations of a
ground water flow model, applied on a typical micro-catchment in
the region are used. The model used is MODFLOW (Harbaugh
2005), which is available as an open-source software (Winston
2006). This three-dimensional distributed physically based model
Short-duration intensities on the days of deluge takes recharge at the land surface as input and simulates variations
Figure 7 shows the 15-min rainfall values recorded at Kudigana in the ground water table and ground water outflow as the
on the 16th of August. It is interesting to note that the rainfall outputs. It uses aquifer parameter values that can be obtained
is nearly continuous on days like this, and the intensities are from field measurements. When coupled with a surface runoff
not extraordinary. Figure 8 shows the intensity–duration curves model, which estimates overland flow and infiltration amounts,
obtained from the data of the 3 days (13th, 15th and 16th) of MODFLOW can be used as a watershed model to simulate runoff
excessive rainfall at the station. The following observations are from the catchment. Ghanshyam (2019) modified this model to
important: facilitate application in variable source area regions (Putty 2010),
in which the recharge zone is a dynamic one. In this type of
(i) The maximum 15-intensity for the days are 60, 80 and 102 applications, the catchment will comprise two parts, the saturated
mm/h, respectively; and the unsaturated. Rainfall on the saturated zone, which usually
(ii) The hourly maximum is about 60 mm/h; lies riparian to the stream channels, will immediately runoff as
(iii) Three-hour maximums are just about 35 mm/h; ‘source area flow’ and rainfall on the unsaturated part, in excess of
(iv) Intensities for durations longer than 10 h are nearly same at initial abstractions, will get infiltrated completely. The infiltration
around 15–25 mm/h. These intensities are not unusually recharges subsurface storage and augments ground water in
high. Very commonly, floods resulting due to overland flow course of time. Mountainous and forested/grassed areas of
are attributed to be the cause for landslides (Parkash 2019) Sahyadris belong to this category of catchments. In the method
and mitigation measures prescribed include provision for developed by Ghanshyam (2019) for applying MODFLOW to suit
draining overland flow properly. Overland flow occurs when the dynamic nature of the recharge zone, information on the

Landslides 18 & (2021) 1605

 Original paper
Table 3 Magnitude and return period of extreme rainfall values for the year 2018 compared with highest ever in the records
Station/data length RF (mm)/return period 1 day 2 days 3 days 5 days 10 days
Madikeri (1960–2018) 2018 RF 184.2 320.6 440.8 681.6 1139.8
T (years) 7.2 9.5 13.0 28.3 68.8
Highest RF 250.8 371.6 462.6 681.6 1139.8
Madapura (1970–2018) 2018 RF 206.0 373.6 487.2 749.6 1066.6
T (years) 28.6 49.3 40.1 74.0 65.6
Highest RF 262.4 386 502.4 749.6 1066.6
Suralabbi (1998–2018) 2018 RF 400.4 778.4 1101.4 1693.4 2826.4
T (years) 18.2 48.7 57.9 65.6 57.9
Highest RF 482 778 1101 1693 2826
Galibeedu (1970–2018) 2018 RF 352 702 932.2 1263 1966.9
T (years) 142.9 369.1 475.4 447.6 561.5
Highest RF 352 702 932.2 1263 1966.9
Sampaje (1960–2018) 2018 RF 177.6 320.6 469.8 729.4 1124.2
T (years) 3.5 6.0 11.5 41.9 62.9
Highest RF 275.6 447.5 608.5 729.8 1124.2
Bhagamandala (1960–2018) 2018 RF 252.0 502.4 667.8 979.0 1397.0
T (years) 2.1 3.5 3.5 4.5 4.2
Highest RF 530 790 1061.8 1452.2 1968.6

variable saturated area has to be input manually to the model in Engineering (N.I.E., Musuru), to which the authors are affiliated.
steps, at the end of each of which the model outputs the ground The model has been applied using daily rainfall data from
water levels and ground water flow. The source area runoff is Galibeedu for 3 years—2011, 2013 and 2018 (‘Progress of the mon-
estimated as equal to the rainfall on the fraction of the catchment soon rain’) and the simulations are studied. A 30-m resolution
which is saturated. The model can be run, for convenience, using DEM is used to obtain the topography of the watershed. The
steps of 5 or 10 days each, when daily rainfall data is used. The catchment is assumed to be made up of two layers vertically, with
model simulates daily runoff and water table level. Further details the total soil thickness varying from 2 m adjoining the streams to
of the model are not furnished here for want of space, but can be 40 m near the ridges. Parameter values used by Ghanshyam (2019)
obtained from the authors. for a nearby catchment, on which the model was applied in order
In the present work, the model has been applied on a micro- to test its performance and found to be reliable, have themselves
catchment of area at 0.53 km2, located in the Kumaradhara Field been used for the present case also. The initial level of the ground
Hydrological Laboratory in Kodagu, a facility being maintained by
the Water Resources Centre of The National Institute of

120.0

100.0
30
Intensity (mm/hr)

15.8.2018
25 80.0 16.8.2018
Rainfall (mm)

20 13.8.2018
60.0
15
40.0
10

5 20.0

0 0.0
1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97
0 200 400 600 800 1000 1200 1400
Interval No.
Duration (min.)

Fig. 7 Fifteen-minute rainfall recordings of days 15th–16th August at Kudigana

(time beginning at 0830 hours on the 15th) Fig. 8 Intensity–duration curves of 3 days of excessive rainfall

1606 Landslides 18 & (2021)

 100
No. of times exceeded
80 15-16 August

60

40

20
15.6.2018 – 9.3% 25.6.2018 – 16.4%
0
102 68 54 40 34 30 26 20 16 12 8 4

Intensity (mm/hr)

Fig. 9 Intensity prevalence—the No. of 15-min intervals with higher intensities

water table is fed in to the model, based on the experience of the

authors and has been assumed to vary between 3 and 10 m below 5.7.2018 – 23.3% 15.7.2018 – 29.6%
the ground level. The simulations of the model are presented in
Fig. 10 and in Table 4. The way the saturated source area expanded
during the monsoon of 2018 is shown in Fig. 10. The elevation and
slope map of the catchment are presented in Fig. 11. Table 4
presents the pattern of growth of saturation during the 3 years
and the rainfall during the corresponding period. It can be ob-
served from these illustrations that the saturation area lies in the
flat valley portions of the catchment during the first few fortnights
after the beginning of the monsoon (assumed to be by 1st of June),
while it grows to rise to higher elevations and cover steeper slopes
during the wetter days. It is observed that by 9th of August 2018,
25.7.2018 – 39.5% 4.8.2018 – 48.8%
the saturation zone had grown to about 54% of the catchment and
covered very steep portions of the land. The rainfall in the 1-week
period following it has been over 1500 mm, an extreme combina-
tion probably unprecedented in the recent history of Kodagu. It
can be argued that such a combination of extreme conditions
resulted in the disastrous landslides of the year. Table 4 shows
that 2013 also saw extreme conditions of saturated areas growing
to over 45% by 9th of August. However, the rainfall by this time of
year is found to have decreased, not having led to disasters of the
kind of 2018. 9.8.2018 – 54.2% 14.8.2018 – 46.1%
Conclusions
According to DST (2016), investigation of known sites is a thrust
area in landslide research. The present work was taken up in order
to investigate the events of 2018 that led to extensive and very deep
landslides in Kodagu in Karnataka. The interferences drawn from
studies on the location of the major landslides, the soil properties
at these sites, rainfall characteristics during the year and a model
analysis can be summarised into the following conclusions:

& Landslides have occurred on all sorts of land, both disturbed 19.8.2018 – 35.3% 24.8.2018 – 30.3%
and undisturbed, implying that anthropogenic activities are
not the primary cause of the landslides; Fig. 10 Variations in the saturated zone in a catchment—the model simulations
& Very deep and porous soils, characterised by high infiltration
rates and a low permeability matrix, are prone to excessive
pore pressures due build-up of the ground water table;

Landslides 18 & (2021) 1607

 Original paper
Table 4 Growth of saturation zone—the model simulations
Date Year
2011 2013 2018
RF (mm) Saturated area (%) RF (mm) Saturated area (%) RF (mm) Saturated area (%)
Up to 15 Jul 1160.7 25.0 2405.9 27.9 2308.1 29.6
25 656.4 29.8 978.3 35.3 760.9 39.5
4 Aug 349.1 37.1 609.8 41.2 331.0 48.8
9 438.9 30.1 198.8 47.1 607.3 54.2
14 194.6 24.7 186.6 37.3 968.0 46.1
19 67.2 18.9 129.6 28.6 840.0 35.3

& The exceptionally wet 10 days between 9th and 17th of August,
which followed a highly wet 2-month period, seems to have
been the factor leading to landslides; 3-day and 5-day extreme
rainfall magnitudes during this period have had return periods Acknowledgements
of around 20 and 50 years, respectively. These extreme rainfall This investigation was taken up as a part of a major project
values at Galibeedu, places near which were the worst affected, sponsored by the Ministry of Earth Sciences (MoES), Govt. of
had return periods of 140 to 2000 years; India. The authors acknowledge with thanks the authorities of
& Despite very high magnitudes of daily rain, overland flow is MoES for funding the project. A part of the rainfall data used in
the study has been made available by the Karnataka Directorate of
not widespread since short-duration intensities are only mod-
erate and infiltration rates are high; Simulations of the water- Economics and Statistics. The subsurface profile data has been
shed model indicate that the growth of saturated zones to very furnished by the Dept. of Mines and Geology, Karnataka. The
steep portions of the valley combined with an unprecedented authors acknowledge with thanks the authorities of these organi-
spell of rain is the most possible triggering factor for the sations for the help.
disastrous events in 2018.
Data availability
The datasets generated during and analysed during the current
study are not publicly available since they have been procured
through intensive field work by the students and will be used by
the next batch of students. However, data are available from the
corresponding author on reasonable request. Daily rainfall data
for Karnataka can be procured on request from Karnataka
Directorate of Economics and Statistics, Bengaluru, India.

References

Abbate A, Longoni L, Ivanov VI, Papini M (2019) Wildfire impacts on slope stability
triggering in mountain areas. Geosciences 2019(9):417
Anbalagan R (1992) Landslide hazard evaluation and zonation mapping in mountainous
terrain. Eng Geol 32(1992):269–277. https://doi.org/10.1016/0013-7952(92)90053-2
Bidie G (1869) On the effects of forest destruction in Coorg. Proc. of the Royal
Geographical Society of London. Vol. 13, No. 2 (1868 -1869). pp. 74-83. Wiley on
behalf of the royal geographical society. http://www.jstor.org/stable/1799393.JSTOR.
Accessed 08-05-2017 05:34 UTC.
Bourgeon G (1989) Explanatory booklet on the reconnaissance soil map of forest area -
Western Karnataka and Goa. French Institute. Pondicherry. India 95 pp.
BRO (2018) Report on visit of BRO team to Kodagu district. Karnataka for restoration of
road connectivity between 28th Aug. 18 to 01st Sep. 18. Border Roads Organisation,
Ministry of Defence. Govt. of India

Fig. 11 The elevation and slope maps for the catchment

1608 Landslides 18 & (2021)

Bruijnzeel LA (2004) Hydrological functions of tropical forests. Not seeing the soil for the Kumar KSS (2005) Geoinformatics in landslide risk assessment and management in parts
trees? Agric Ecosyst Environ 104:185–228. https://doi.org/10.1016/j.agee.2004.01.015 of Western Ghats, Central Kerala, South India. Ph.D. Thesis, Indian Inst. of Tech.,
Burt TP, McDonnell JJ (2015) Whither field hydrology? The need for discovery science Bombay
and outrageous hydrological hypotheses. Water Resour Res 51:5919–5928. https:// Kuriakose SL, Sankar G, Muraleedharan C (2008) History of landslide susceptibility and a
doi.org/10.1002/2014WR016839 chronology of landslide prone areas in the Western Ghats of Kerala India. Environ
Chaitra, Putty MRY, Prasanna HS (2015) Subsurface drainage and storage properties in Geol 57(7):1553–1568. https://doi.org/10.1007/s00254-008-1431-9
the Western Ghats–a study in the basin of Netravati. Int. Conf. ICWRCOE - 2015. Li WC, Dai FC, Wei YQ, Wang ML, Min H, Lee LM (2015) Implication of subsurface flow on
March, 2015 NITK, Surathkal, Karnataka. https://doi.org/10.1016/j.aqpro.2015.02.080 rainfall-induced landslide: a case study. Landslides 13:1109–1123
Chau LN, Tien PV, Do TN (2019) Deep-seated rainfall-induced landslides on a new Maneesha VR, Vasudevan N (2012) The deployment of deep-earth sensor probes for
expressway: a case study in Vietnam. Landslides:11/21. https://doi.org/10.1007/ landslide detection. Landslides 9:457–474. https://doi.org/10.1007/s10346-011-0300-x
s10346-019-01293-6 Martha T, Roy P, Govindharaj KB, Kumar VK, Diwakar PG, Dadhwal VK (2014) Landslides
Chawla A, Chawla S, Pasupuleti S, Rao ACS, Sarkar K, Dwivedi R (2018) Landslide triggered by the June 2013 extreme rainfall event in parts of Uttarakhand state, India.
susceptibility mapping in Darjeeling Himalayas. India. Adv Civ Eng 2018. https:// Landslides 12:135–146. https://doi.org/10.1007/s10346-014-0540-7
doi.org/10.1155/2018/6416492 Maruthi KV, Chandran KV (2013) Landslide Hazard Zonation of Virajpet-Madikere-Sulya
Darshan R (2016) Some studies on hydrological properties of soils in the upland areas of Road Section Coorg District Karnataka. Unpub. GSI Report of FS 2012- 2013
the Gundia Basin. Unpub. M. Tech. dissertation, The National Institute of Engg., Mishra AK, Pradhan UK, Singh LP (2010) A report on the Landslide Hazard Zonation of
Mysuru, India Madikere town on macroscale Coorg district Karnataka. Unpub. GSI Report of FS 2007-
Dept. of Sc. and Tech. (DST) (2016) Thrust areas under landslides in India, Ministry of Sc. 2009
and Tech., Govt. of India Moriwaki H, Inokuchi T, Hattanji T et al (2004) Failure processes in a full-scale landslide
Fonseca LDAM, Lani JL, Santos GRD, Ferreira WPM, Santos AMRT (2017) Variability in soil experiment using a rainfall simulator. Landslides 1:277–288. https://doi.org/10.1007/
physical properties in landslide - prone areas. Acta Scientiarum. Agronomy 39(1):109– s10346-004-0034-0
118. https://doi.org/10.4025/actasciagron.v39i1.30561 Nitish MG (2019) A study on recent landslides in Kodagu. Unpub. M. Tech. dissertation,
Forbes K, Broadhead J (2011) Forests and landslides - the role of trees and forests in the N.I.E., Mysuru-570008
prevention of landslides and rehabilitation of landslide-affected areas in Asia. Food Parkash S (2019) Landslide preparedness guidelines for safety of buildings on slopes
and Agriculture Org. of the UN, Reg.Office for Asia and the Pacific, Bangkok 2011, 42 published by National NIDM, Ministry of Home Affairs, Gov. of India New Delhi-
pp 110001 India pages 80. ISBN No.: 978-93-82571-30-8
Ghanshyam, Giri (2019) A critical study on the application of MODFLOW in Western Ghat Pascal JP (1984) Wet evergreen forests of the Western Ghats. French Inst., Pondicherry, p
watersheds. Unpub. M.Tech. dissertation. The National Institute of Engg., Mysuru, 27
India Putty MRY (2006) Salient features of the hydrology of the Western Ghats in Karnataka.
Giri P, Ng K, Phillips W (2018) Laboratory simulation to understand translational soil In: Krishnaswamy J, Lele S, Jayakumar R (eds) Hydrology and watershed services in
slides and establish movement criteria using wireless IMU sensors. Landslides the Western Ghats of India. Tata-McGraw Hill Co. Ltd.
15:2437–2447. https://doi.org/10.1007/s10346-018-1055-4 Putty MRY (2010) Principles of hydrology. IK International Pub, New Delhi
GSI (2018a) A note on the preliminary post disaster investigation of landslides occurred Putty MRY, Prasad R (2000) Runoff process in head water catchments – an experiment
around Madikeri Kodagu district Karnataka. Geological Survey of India. State Unit: study in Western Ghats South India. J Hydrol (Netherlands) 235:63–70
Karnataka and Goa Bangalore August 2018 Putty MRY, Swamy BKK, Prasad R (1997) A study on infiltration rates in a very wetzone of
GSI (2018b) Report on post disaster investigation of landslide occurrences along Madikeri the Western Ghats. Hydrol J IAH 20(1.4):1–17
- Makutta and Madikeri – Koyanad - Sulya road sections, Kodugu District, Karnataka. Ramachandra TV, Subashchandran MD, Ashisar AH (2009) Landslides at Karwar, October
Geological Survey of India. State Unit: Karnataka and Goa Bangalore August 2018 2009: causes and remedial measures. ENVIS technical report: 33, November 2009,
Hamilton LS (2008) Forests and water: a thematic study prepared in the framework of CES, I.I.Sc., Bangalore – 560 012. http://ces.iisc.ernet.in/energy. Accessed 24 Mar 2020
the global forests resources assessment 2005. FAO Rome.sidalc.net Ramachandra TV, Subashchandran MD, Joshi NV, Pallav J, Uttam K, Aithal BH, Vishnu M
Harbaugh AW (2005) MODFLOW-2005, the U.S. Geological Survey modular ground-water (2012) Landslide susceptible zone mapping in Uttara Kannada Central Western Ghats.
model—the ground-water flow process: U.S. Geological Survey Techniques and ENVIS Technical Report 28. Energy and Wetlands Research Group. Centre for Ecolog-
Methods, 6-A16 ical Sciences Indian Inst. of Sc. Bengaluru 560012
Huntley D, Bobrowsky P, Hendry M et al (2019) Application of multi-dimensional Reder A, Pagano L, Picarelli L et al (2017) (2017) The role of the lowermost boundary
electrical resistivity tomography datasets to investigate a very slow-moving landslide conditions in the hydrological response of shallow sloping covers. Landslides 14:861–
near Ashcroft British Columbia Canada. Landslides 16:1033–1042. https://doi.org/ 873. https://doi.org/10.1007/s10346-016-0753-z
10.1007/s10346-019-01147-1 Richter GR (1870) Gazetteer of Coorg. BR Pub. Corp., Delhi
Ibsen M, Casagl N (2004) Rainfall patterns and related landslide incidence in the Sajinkumar KS, Anbazhagan S, Pradeepkumar AP, Rani VR (2011) Weathering and
Porretta-Vergato region Italy. Landslides 1:143–150. https://doi.org/10.1007/ landslide occurrences in parts of Western Ghats, Kerala. J Geol Soc India 78:249–257
s10346-004-0018-0 Sankar G, Varma AK, Sekhar LK, Rajan PP, Deepa C, Eldhose K (2017) Land disturbances
ISRO (2018) Home Disaster Management Support Landslides, Dept. of Space, Indian due to Soil piping in the Western Ghats region of Kerala. Proc. Nat. WS and field trng
Space Research Org. https://www.isro.gov.in/landslides-karnataka. Accessed 15 Mar on Land disturbances due to soil piping in Western Ghats, July 5-9, 2017, NCESS,
2020 Thiruvanantapuram, Kerala
Jackson IJ (1988) Daily rainfall over Northern Australia. Deviations from the World Scaioni M, Longoni L, Melillo V, Papini M (2014) Remote sensing for landslide investi-
pattern. J Climatol 8:463–476. https://doi.org/10.1002/joc.3370080503 gations an overview of recent achievements and perspectives. Remote Sens
Jain R, Madhusoodhanan S, Nair AS, Azmal M, Rani N (2019) Disaster analysis of Kodagu 6(10):9600–9652. https://doi.org/10.3390/rs6109600
District using Geomatics Technology. Int J Eng Res Senthilkumar V, Chandrasekaran SS, Maji VB (2017) Geotechnical characterization and
Technol 08(2019/06):IJERTV8IS060634 analysis of rainfall—induced 2009 landslide at Marappalam area of Nilgiris district,
Jones JAA (1994) Soil piping and its hydrogeomorphic function. Cuaternario Y Tamil Nadu state, India. Landslides 14:1803–1814. https://doi.org/10.1007/s10346-
Geomorfopia 8(3-4):77–102 017-0839-2
Kanungo D, Arora MK, Sarkar S, Gupta RP (2009) Landslide susceptibility zonation (LSZ) Sharma LP, Patel N, Debnath P, Ghose MK (2012) Assessing landslide vulnerability from
mapping - a review. J S Asia Disaster Stud 2:81–105 soil characteristics a GIS-based analysis. Arab J Geosci 5(4):789–796. https://doi.org/
Khang D, Kyoji S et al (2019) Recent rainfall-induced rapid and long-traveling landslide 10.1007/s12517-010-0272-5
on 17 May 2016 in Aranayaka Kagelle District Sri Lanka. Landslides 16:155–164. Star of Mysore (SOM) (2018a) Deforestation heavy rainfall caused destruction in Kodagu.
https://doi.org/10.1007/s10346-018-1089-7 Report, September 14 2018
Kodagu District Disaster Management Plan 2017-2018, Govt. of Karnataka. Star of Mysore (SOM) (2018b) Geologists pinpoint extensive land damage for Kodagu
www.kodagu.nic.in disaster. Report, September 27, 2018

Landslides 18 & (2021) 1609

 Original paper
The Hindu (2018) “Expert calls for landslide hazard zonation mapping” – Report by Zhenlei W, Sun H-y, Xu H-d, Wang D-f (2019) The effectiveness of a drainage tunnel in
Krishna Kumar. Mysuru. August 20 2018 increasing the rainfall threshold of a deep-seated landslide. Landslides 16:1731–1744.
Tsukamoto Y, Ohta T, Nogushi H (1982) Hydrological and geomorphological studies of https://doi.org/10.1007/s10346-019-01241-4
debris slides on forested hill slopes in Japan. Int Assoc Sci Publ 137:89–98
Van Asch TWJ, Buma J, Van Beek LPH (1999) A view on some hydrological triggering
M. R. Y. Putty ()) : B. N. Prithviraj : P. N. Kumar : M. G. Nithish :
systems in landslides. Geomorphology 30:25–32 PII: S0169- 555X_99.00042-2
. Giri : P. N. Chandramouli
Walker LR, Shiels AB (2013) Physical causes and consequences for landslide Ecology.
USDA National Wildlife Research Center Staff Publications. 1640. https://
Department of Civil Engineering,
digitalcommons.unl.edu/icwdm_usdanwrc/1640. Accessed 24 Mar 2020
The National Institute of Engineering,
Winston RB (2006) GoPhast: a graphical user interface for PHAST: U.S. Geological Survey
Mysuru, Karnataka 570008, India
Techniques and methods 6-A20, 98 p.
Email: puttyyadupathi@gmail.com; puttymys@nie.ac.in
Zhao W, Wang R, Liu X, Ju N, Xie M (2020) Field survey of a catastrophic high-speed
long-runout landslide in Jichang Town, Shuicheng County Guizhou China on July 23,
2019. Landslides 17:1415–1427. https://doi.org/10.1007/s10346-020-01380-z

1610 Landslides 18 & (2021)