Printer Friendly

Applying a GIS slope-stability model to site-specific landslide prevention in Honduras. (Research).

Keywords: Central America, Hurricane Mitch, slope failure, spatial modeling

Steep lands (slope >30%) comprise 34% of all land area in the tropics. These lands are susceptible to degradation by soil erosion caused by gradient, rainfall erosivity, and-in humid regions of the tropics-intense soil weathering (Lal, 1999). When natural vegetation on humid steep lands is cleared for timber harvest or for agriculture, this inherent susceptibility translates into excess soil loss and increased incidence of mass soil movement, including debris slides, mudflows, and landslides. Mass soil movement is frequently the dominant denudational process in steep humid areas (lida, 1999; Nagle et al., 1999; Pla, 1997), causing catastrophic damage to lives and property (Chung et al., 1995) and contributing to progressive problems of river degradation, siltation of reservoirs, and contamination of drinking water (Cruz and Reyes, 2000; Nagle et al., 1999; Pla, 1997).

The economic cost of these problems is greatest in developed countries, but 95% of all landslide-related deaths occur in the developing world (Hansen, 1984). Even within developing nations, the human impact is unevenly distributed, as the economically or socially marginalized are forced to occupy the most hazardous slopes in rural and urban areas (Alexander, 1995). Population growth and degradation of traditional agricultural lands have intensified pressure on these fragile areas, increasing total human vulnerability to slope failure (Sheng, 1989).

Hurricane Mitch (October 1998) provided devastating evidence of human vulnerability to hillside instability. The hurricane was the largest storm to strike the Atlantic Basin in more than 10 years (Hellin and Haigh, 1999) and the deadliest in more than 200 years (McCown et al., 1998). About 11,000 people in Central America and the Caribbean lost their lives, thousands more were left homeless, and the lasting geomorphic effects of the storm have increased susceptibility to floods and landslides throughout the region (IADB, 1999).

Landslide management. Catastrophic landslides leave farmers and land managers of Central America with a considerable challenge: to manage against slope failure--and associated mudflows and inundation--at a time when pressure on fragile steep lands is at its greatest. Local and international agencies have worked on this problem in the years since Mitch, exerting considerable effort to identify management strategies that will limit surface erosion and prevent mass soil movement (e.g. World Neighbors, 2000; Cruz and Reyes, 2000). Reforestation and improved road drainage, both established slope-stabilization strategies (e.g. Bergin et al., 1995; Sidle et al., 1985; Royster, 1979), have received important attention (Cruz and Reyes, 2000; IADB, 1999), but there is still a need to identify agricultural practices that reduce vulnerability to slope failure. Given the rapid demographic expansion found in much of the Mitch-affected region, it simply won't be possible to set aside all vulnerable slopes for protection; th erefore, methods for sustainable production on steep lands are required (Lal, 1994; El-Ashry, 1988).

A soil conservation survey conducted in Honduras, Nicaragua, and Guatemala in the wake of Hurricane Mitch failed to find any beneficial effect of "agroecological" farming on landslide resistance during the storm (World Neighbors, 2000). Use of agroforestry, contour cropping, physical and vegetative erosion barriers, or integrated weed management was found to reduce soil degradation and surface erosion on farms affected by Mitch, but these techniques did not correlate with a reduction in landslides. While it is possible that soil conservation farming truly has no effect on slope stability, the experience of geotechnical engineers and landslide modelers indicates that land management targeted to soil or site-specific conditions can reduce the probability of slope failure (Montgomery et al., 2000; Collison et al., 1995; Royster, 1979).

Unfortunately, the biophysical complexities of the problem are nearly as great as the social and economic. Surface erosion of soil is influenced by slope form (length, shape, and topography), rainfall (intensity, duration, and frequency), soil properties (granulometry, structure, aggregate stability, and strength), hydrology (antecedent water content, hydraulic conductivity, and slope drainage), and vegetation characteristics (density, height, and management). Slope stability is influenced by the same factors and also depends on subsurface pore pressures, the distribution and strength of roots, localized weathering of surficial bedrock, and hillside geohydrology. The number of variables involved makes it difficult to characterize the influence of land management on landslide susceptibility and even more difficult to predict how changes in management will impact the current state (El-Swaify and Fownes, 1992). It is hardly surprising that efforts to conserve soil on tropical steep lands have often failed becaus e of recommendation of inappropriate practices (Shaxson, 1988).

In recent years, the field of physical landslide modeling has made excellent progress in dealing with the complexities of slope failure. Applications of simplified slope-stability models have proved effective as descriptive and predictive tools in temperate zones, allowing for rapid stability assessment over a wide area (e.g. Miles et al., 2000; Pack et al., 1998; Wu and Sidle, 1995; Montgomery and Dietrich, 1994; Jibson, 1993). Few models have been developed expressly for stability analysis in the tropics, but applications of simplified slope-stability models have performed reasonably well (Divakarla and Macari, 1998; Terlien et al., 1995) despite the influence of localized weathering and subsurface flow on tropical slope stability (Simon et al., 1990). It is rare, however, for landslide prediction models to be fully utilized as prescriptive tools for land management. Most provide the end user with a spatially distributed stability index but offer no indication of the relative significance of modeled process es that drive the calculation. Knowledge of a stability index alone may be useful for planning timber harvests (e.g. Montgomery et al., 2000) or for citing infrastructure (e.g. Carrara et al., 1995), but it is not helpful for steep-land farmers for whom non-use and relocation are not viable options. If physical landslide models are to contribute to slope stabilization in tropical, agricultural watersheds, then a driving cause for instability must be identified, along with the probability that a slope will fail.

Importance of causal information is born out by physical slope-stability models that do consider the influence of dynamic hydrology and plant growth. These models are computationally intensive and, to our knowledge, have only been applied to engineered or "characteristic" slope formations (Pla, 1997; Duncan, 1996; Collison et al., 1995; Anderson et al., 1988). Nonetheless, results of these modeling exercises are relevant to field-level management. Collison et al. (1995) found that planting trees on an engineered embankment only enhances the embankment's stability if soil hydraulic conductivity is relatively high. For deep-soiled embankments with low hydraulic conductivity, planting trees was, in fact, detrimental to stability. Macropore flow along tree roots increased permeability, leading to elevated pore pressures in and below the rooting zone, while roots themselves did not penetrate deeply enough to anchor against deep-seated slope failure. Under such conditions, it is preferable to vegetate the embankmen t with grass, or some similar shallow-rooted ground cover, that sheds water off the saturated portion of the slope. Pla (1997), modeling a slope characteristic of the Western Andes of Venezuela, found that simple reforestation of a degraded pasture was not sufficient to prevent landslides--hydraulic conductivity was so low in the argillic horizon of Ultisols that drainage ditches were required to prevent saturation-induced translational failure. Both these studies indicate that tree planting is not the solution to slope instability in all cases. Water management can be more important than physical reinforcement for weathered and frequently saturated soils, and drainage ditches, or water-shedding ground cover, may offer better slope stabilization than trees. Information on the soils and hydrology of a landslide-prone area is therefore necessary for the development of appropriate management recommendations; simply reporting the probability of failure is not enough.

The objectives of this study were (1) to apply and evaluate a simplified slope-stability model to an area in Central Honduras that was severely affected by Hurricane Mitch, and (2) to develop landslide-susceptibility maps categorized according to local slope gradient ([alpha]) and relative wetness (W), the variables that most influence slope stability. Differentiation of susceptible zones into slope-wetness ([[alpha].sup.*] W) classes is intended to assist land management at the farm and community level.

Methods and Materials

Study area. The study area comprises three contiguous watersheds in the central highlands of Honduras (Figure 1), those of the Rio Salitroso, the Rio Frio, and the headwaters of the Rio Quirima. All three rivers drain into Embalse Francisco Morazan--the reservoir of the largest hydroelectric dam in the nation, which provides Honduras with the majority of its electricity (Mangurian, 1997). Rapid siltation of the dam's reservoir is a serious concern to the government of Honduras and to international funding agencies (COHDEFOR, 1998; IADB, 1999). The study area lies within the municipality of La Libertad, Districto Comayagua (N 14[degrees] 47', W 87[degrees] 35') and covers an area of 46.1 [km.sup.2] (120 [mi.sup.2]). Land use in the project area is predominantly agricultural, with coffee cultivation (Coffea arabica, L.) dominating upland areas and lowlands divided between cattle pasture and the production of maize and beans (Zea mays, L., Phaseolus vulgaris, L., Phascolus lunatus, L.). Most farms are between 1 and 2 ha (2.5 and 5 ac) in size and are cultivated by the landowners (COHDEFOR, 1998).

The central Honduran highlands are semi-humid tropical. They receive an average of 1,900 mm (74.8 in) of precipitation annually, concentrated in a moderate rainy season from May through July and a heavy rainy season from September through November. Terrain is steeply dissected and soils are residual, of volcanic and sedimentary origin. Hurricane Mitch inflicted severe damage to the study area, triggering significant debris flows, mudslides, surface erosion, lowland flooding, and the consequent destruction of homes and farms (Cruz and Reyes, 2000).

Data collection. Elevation contours were hand-digitized in ArcView GIS (ESRI Inc., Redlands, CA) from 1:50,000 topographic basemaps (NIMA, Fairfax, VA). Contours were used to interpolate a raster Digital Elevation Model (DEM) at the 5 m scale (after Zhang and Montgomery, 1994). Slope and flow path were derived from the DEM.

A total of 63 soil samples were collected from the surface and 0.75 m (2.5 ft) depth (B-horizon). Sample sites were chosen to cover the range of observed soil and land-cover types; aspect, gradient, and landslide history were not considered in the selection of sample locations. Bulk density and saturated shear strength ([[TAU].sub.sat]) (Torvane Shear Strength Device, ELE International, Lake Bluff, IL) were measured and recorded for each sample. The saturated strength represents the most conservative strength estimate for the sandy-clay soils that dominate the research area; soil strength properties can vary significantly with degree of saturation (Fredlund, 2000), but the conservative estimate has proved successful in earlier landslide studies performed at the watershed scale (e.g. Iverson, 2000; Divakarla and Macan, 1998; Montgomery and Dietrich, 1994). Angle of soil friction for the samples ([phi]) was estimated from information on soil texture and porosity (Holtz and Kovacs, 1981), and saturated hydraulic conductivity ([K.sub.sat]) was evaluated with the falling-head method (Reynolds, 1993) for 76 mm (3 in) inside diameter and 76 mm (3 in) height undisturbed soil cores, based on the average of three trials per sample. Soil thickness soil above bedrock (D) was measured at 50 points (road cuts and sample pits).

Three major soils were identified, and their distribution was mapped through correlation with surface lithology. Information on lithology was obtained from 1:500,000 U.S. Geologic Survey (USGS) maps for Honduras that were refined through field survey. This refinement was necessary because USGS geologic data were disaggregated to match the scale of our DEM. Disaggregation is known to introduce uncertainty to a spatial model (Heuvelink and Pebesma, 1999), but, when the process is justified with field survey, it allows the modeler to present more detailed, intuitive results for use in prediction and management (Miles et al., 2000; Montgomery and Dietrich, 1994). Lithologic types include limestone deposits (Valle de Angeles Group), pyroclastic rock (Matagalpa Formation) and calcareous shale (Yojoa Group). Measured soil properties were grouped by lithologic type to provide calibration parameters for the stability model.

No reliable data on rainfall from Hurricane Mitch were available for the study area, and, because of the highly destructive nature of the storm, data from rainfall gauges throughout Honduras were of questionable quality (Hellin and Haigh 1999). Maximum 30-minute intensities of 71 mm [hr.sup.-1] (2.8 in [hr.sup.-1]) were reported from the south of the country (Helm and Haigh, 1999), while a sustained intensity of nearly 107 mm [hr.sup.-1] (4.2 in [hr.sup.-1]) was recorded over one six-hour period in the north (McCown et al., 1998). Based on these reports, we used a steady-state recharge rate of 76 mm [hr.sup.-1] (3 in [hr.sup.-1]) for modeling purposes, considered conservative for maximum, multiple-hour, sustained rainfall intensity in central Honduras.

An inventory of recent landslides in the study area was established through a field survey in September 2000 and the interpretation of 1:40,000 black and white aerial photographs (USGS, Washington, D.C.) taken in March 2000. All landslides were mapped at the head scarp of the slope failure (after Pack et al. 1998). Point mapping was considered to be the most appropriate inventory method because (1) dense vegetation in the tropics introduces considerable uncertainty to slide-area maps and (2) volcanic soils may liquefy when disturbed, leading to mudslides that are only a secondary consequence of initial slope failure (Terlien et al., 1995).

Model implementation. Following Montgomery and Dietrich (1994), a physically based slope-stability model was used in this study. Physical models allow for relatively fine-scale hazard mapping (Pack et al., 1998) and tend to be less site-specific than multi-variate statistical analyses (Montgomery and Dietrich, 1994). The model most commonly applied in the GIS environment is the infinite slope-stability model (e.g. Hammond et al., 1992; Montgomery and Dietrich, 1994; Wu and Sidle, 1995; Terlien et al., 1995; Pack et al., 1998). This model evaluates slope stability based on an estimated potential for failure, expressed as (Hammond et al., 1992): where FS is the "factor of safety" for slope

FS = [c.sub.r] + [c.sub.s] + [cos.sup.2][theta] [[rho],g (D - Dw) + ([rho],g - [[rho].sub.w]g) Dw] tan [phi]/D [rho],g sin [phi] cos [phi] (1)

stability. (Values < 1 represent failure conditions.) Parameters [c.sub.r] and [c.sub.s] are the cohesion due to roots and soil, respectively ([ML.sup.-1][T.sup.-2]), [theta] is the local slope gradient in degrees, [rho]s is the saturated bulk density of the soil ([ML.sup.-3]), [rho]w is the unit density of water ([ML.sup.-3]), D is the soil thickness above the failure plane (L), Dw is the saturated soil thickness above the failure plane (L), [phi] is the soil angle of friction in degrees, and g is the acceleration due to gravity ([LT.sup.-2]).

Pack et al. (1998) adapted this equation in the Stability Index Mapping (SINMAP) model. SINMAP estimates shallow translational landslides by combining the infinite slope-stability equation with a steady-state hydrology model that calculates depth of saturation based on water recharge rate and soil transmissivity of water. In numerous applications, SINMAP has proven to be an effective tool for the rapid identification of potentially unstable zones in large land areas (Pack et al., 1998). Although shallow translational landslides were not the only type of slope failure identified in the study area, the infinite slope model was deemed appropriate because (1) the algorithm tends to be conservative for situations of deep-seated failure (Terlien et al., 1995), and (2) rotational failures and mudslides in the study area generally initiate as shallow debris flows.

Slope-wetness classification. Each grid cell was categorized according to relative wetness (W) and average slope ([alpha]). W was obtained from the SINMAP hydrology module and, for the deterministic case, represented the fraction of soil thickness predicted to be saturated under modeled conditions:

W = (R * A/T sin [alpha], 1), (2)

where R and T are rainfall and transmissivity for each grid cell and A is the specific catchment area (upslope area per unit contour length; [L.sup.2] [L.sup.-1]). Values of W range from 0 to 1, and W=1 indicates complete saturation. Slope ([alpha]) was derived directly from the 5 m DEM and ranged from 0[degrees] to 53[degrees]. Four categories were used to characterize slope-wetness ([[alpha].sup.*] W) variability in the study area: Class A, for grid cells in "rolling terrain" ([alpha] < 30[degrees]) that are "moist" (W < 0.5) under simulation conditions; Class B, for cells in rolling terrain that are wet in model simulation ([alpha] < 30[degrees], W [greater than or equal to] 0.5); Class C, for moist cells in "steep terrain" ([alpha] [greater than or equal to] 30[degrees], W < 0.5); and Class D, for wet, steep terrain cells ([alpha] [greater than or equal to] 30[degrees], W [greater than or equal to] 0.5). The choice of 30[degrees] as a cutoff for steep terrain is consistent with the FAO definition of steep lands (Lal, 1999).

Total and unstable land areas were tabulated by [[alpha].sup.*] W class and lithologic group. Observed slide systems were also categorized by lithology and [[alpha].sup.*] W of the surrounding grid cell, and significance of the lithology by [[alpha].sup.*] W class interaction was tested with the [chi square] statistic. Finally, a map was generated that indicates stability prediction and [[alpha].sup.*] W class for each grid cell in the study area.

Results and Discussion

Soil strength and hydrology were variable within and between lithologic groups (Table 1). Soils overlying Yojoa Group rock were deep and well-drained (D = 3.0 m [10 ft], [K.sub.sat] = [1.06.sup.*][10.sup.-1] [mm*s.sup.-1] [[4.17.sup.*][10.sup.-3] in [s.sup.-1]]). These soils were also the coarsest found in the study area and had the greatest mean [phi] (31.8[degrees]). Soils of the Valle de Angeles group had greater clay content, were relatively shallow ([phi] = 29.3[degrees], D = 1.0 m [3.3ft), and were moderately well-drained with a lower average [K.sub.sat] (3.02*[10.sup.-2] mm*[s.sup.-1] [1.19*[10.sup.-3] in [s.sup.-1]). Soils of the Matagalpa Formation were deep (D = 3.3 m [11 ft]), highly weathered, and clay-rich. These deposits had the lowest [K.sub.sat] (1.86*[10.sup.-2] mm*[s.sup.-1] [7.32*[10.sup.-4] in [s.sup.-1]) and [phi] (28.6[degrees]), but exhibited cohesion under nonsaturated conditions. Landslide density was greatest for the Matagalpa Group, at 3.33 landslide systems per square kilometer (8.66 per square mile).

The SINMAP approach to instability modeling was quite successful in describing slope failure in the study area, identifying 75% of the 81 observed slide systems as sites of expected failure during Mitch. While the SINMAP-defined threshold of instability is sensitive to a number of soil parameters (notably [K.sub.sat] and [phi]; Zaitchik et al., in press), relative stability predictions are governed primarily by [alpha] and W (Pack et al., 1998). Local gradient [alpha] was derived from the DEM, while relative wetness W takes into account [alpha], contributing area, water recharge rate, and hydrology. It is useful to examine the relative and absolute values of [alpha] and W for each soil type and zone of predicted instability.

Only 10.2% of the total study area (4.7 [km.sup.2] [12.2 [mi.sup.2]) was classified as steep terrain (Classes C and D). This small area accounted for 45.2% of all areas predicted to be unstable and 61.7% of all slides observed from Hurricane Mitch (Table 2). It is important to note that the majority of steep-slope failures occurred in Class C areas, where W was < 0.5. These landslides probably occurred under partially saturated conditions and can be referred to as "slope-driven" failures that generally occur in soils that are moderately drained to well-drained.

Wet, rolling terrain (Class B) accounted for 28.2% of the study area (11.9 [km.sup.2]) but 42.4% of areas predicted to be unstable. Class B slopes were not extremely steep ([alpha] < 30[degrees]) but became saturated or submerged under Hurricane Mitch conditions. Class B soils are found in low-lying terrain, along the borders of streams and rivers, and in zones of topographic convergence in low [K.sub.sat] soils. Only 22.2% of all slide systems were found in Class B areas, but several of these triggered deep, extensive mudflows typical of saturated clay soils (Terlien, 1995). In one case, a set of failures on Class B terrain gave rise to a 6 km long mudflow that damaged or put at risk 50 farms (Cruz and Reyes, 2000); this was by far the largest soil-failure event in the study area. So, while Class C failures accounted for the majority of landslides during Hurricane Mitch, the "saturation-driven" failures observed in Class B areas are cause for concern. Additionally, slope stabilization techniques that are generally effective against slope-driven landslides, including tree planting and contour cropping, can actually reduce the stability of slopes prone to deep, saturation-driven failure (Collison et al., 1995; Simon et al., 1990; Royster, 1979). Slope-stability management must involve an array of diverse, site-targeted methodologies.

There was a marked relationship between surface lithology and failure process in the study area. For the coarse-textured Yojoa Group, the majority of predicted and observed failures occurred on steeper slopes (Classes C and D), with few Class B failures observed (Table 2). Some small-debris slides were also found on Class A slopes, but no large failure systems were initiated on rolling terrain. For the Valle de Angeles Group, however, the majority of the area was identified as Class B. Failure under these conditions was driven by rapid surface saturation that compromised the soil matrix even on relatively gentle slopes. The Matagalpa Formation was associated with both slope- and saturation-driven failure. These volcanic deposits are formed on steep slopes in the headwaters of the study area and are kept stable, in part, by soil cohesion, which is lost when positive pore pressures build. Large masses of soil become plastic, or even liquid, resulting in shallow debris flows in steep areas and deep failures and flows in zones of topographic convergence (Terlien et al., 1995). The lithologic group by [alpha]*W class interaction was significant for observed slope failures ([chi square] = 17.9, p = 0.0065), suggesting that there is potential for a soil-specific approach to landslide management.

Considering the diverse nature of slope failure within this study area, it is not surprising that the post-Mitch farm survey (World Neighbors, 2000) failed to identify a link between agroecology and landslide prevention at the regional scale. The surveyors' definition of agroecology included practices as biophysically disparate as, for example, agroforestry and live mulch. Agroforestry may stabilize a shallow soil, where tree roots reach to bedrock, or a well-drained cohesionless soil (Montgomery et al., 2000), such as those found in Class C areas of the Yojoa Group. The increase in depth and rate of infiltration associated with new trees can, however, have a negative impact on stability when a potential failure plane exists below the rooting zone (Collison et al., 1995). Conventional recommendations that involve terracing and tree planting for slope stabilization could lead to destabilization on deep, clay-rich soils of the Matagalpa Formation. Live mulch, on the other hand, could be an excellent tool for la ndslide prevention on a moderately sloped, poorly drained site (i.e. Class B). The shallow roots of the mulch will not reinforce soil along potential planes of failure, but certain mulches can shed water effectively and, when oriented parallel to the slope gradient, may act as thatched roofs (e.g. Triomphe, 1996). Presumably, this will reduce soil saturation and a buildup of positive pore pressures while providing protection from surface erosion.

The [alpha]*W classification approach may be applied to land management at the farm and hillside scale. Two examples are shown in Figure 2. Where a conventional slope stability map identifies both Site 1 and Site 2 as areas of potential instability, the [alpha]* W map indicates that the instability in Site 1, located on a steep Yojoa Group slope, is fairly well drained in the backslope. In this area, integration of deep-rooted woody species into the agricultural system could enhance effective soil strength. The bottom slope, however (located in the South Central area of the map), is both steep and prone to saturation. The site is near a stream channel, so physical drainage is a viable option. Filter strips were also recommended for the banks of the stream channel (Cruz and Reyes, 2000). Site 2, on the other hand, represents a zone of topographic convergence in deep Matagalpa Formation material. Gently graded drainage trenches have been dug in existing landslide scars, a nearby road-cut is recommended to be gr aded back, and vegetation with local grasses is the favored biophysical stabilization method. Water management in this case should be prioritized over soil reinforcement.

Preferred management practices for the four W classes are listed in Table 3. These are proposed based in part on the work of other researchers (Keller and Sherar, 2000; Pla, 1997; Collison et al., 1995; Howell et al., 1991), and are presented as "rules of thumb" to be refined and modified with future field experience.

Summary and Conclusion

Spatially distributed slope-stability models have provided a tool for rapid characterization of landslide susceptibility in large land areas. For land managers concerned with small-scale forest clearing or siting of infrastructure, information on relative susceptibility is generally sufficient--the management decision is one of use or non-use, and low-risk sites are preferable to high-risk sites. For tropical smallholders, however, the biophysical conditions that drive slope instability are as important as relative landslide risk. Few farmers are in a position to relocate their homes or fields, so management decisions involve modification, not elimination, of current land-use practices.

In the absence of human disturbance or geologic faults, relative stability of a slope is controlled primarily by local slope gradient and degree of saturation. The steady-state hydrology module of a model such as SINMAP can, when properly calibrated, provide estimates of the absolute and relative saturation level of soils in the landscape. This information, combined with slope gradient data from a DEM, allows us to define [alpha]* W classes of landslide risk and to identify soils and topographic formations associated with each class. This process requires no more data than the slope-stability model itself, yet it provides stakeholders with new and meaningful information for site-specific landslide management.

A further advantage is that slopes can be classified according to gradient and wetness through field survey (Table 3) when model data is not available; this means that management principles based on [alpha]* W classification can be applied in an approximate manner for slope-stabilization efforts at similar sites around the region. It is expected that the introduction of a slope-wetness component to management discussions will make the exchange of ideas between farmers, agencies, and land planners more fruitful and ultimately help to reduce human vulnerability to catastrophic slope failure.
Table 1

Mechanical and hydrologic parameters used in the slope stability model.

Parameter                  Rock group              n   Mean

[[rho].sub.sat]            Yojoa group             8   1.83
(Mg [m.sup.-3])            Valle de Angeles group  8   1.69
                           Matagalpa formation     14  1.74

[phi](*)                   Yojoa group             8   31.8
                           Valle de Angeles group  8   29.3
                           Matagalpa formation     14  28.7

[[[tau].sub.sat].sup.kPa]  Yojoa group             8    9.9
                           Valle de Angeles group  8    8.9
                           Matagalpa formation     14   6.2

log [K.sub.sat]            Yojoa group             8   -3.97
(m [s.sup.-1])             Valle de Angeles group  8   -4.52
                           Matagalpa formation     14  -4.73

D (m)                      Yojoa group             8     3
                           Valle de Angeles group  8     1
                           Matagalpa formation     14  3.25

Parameter                  Rock group                 Range

[[rho].sub.sat]            Yojoa group              1.80 - 1.86
(Mg [m.sup.-3])            Valle de Angeles group   1.63 - 1.72
                           Matagalpa formation      1.67 - 1.85

[phi](*)                   Yojoa group              35.1 - 28.6
                           Valle de Angeles group   30.4 - 29.1
                           Matagalpa formation      25.6 - 31.7

[[[tau].sub.sat].sup.kPa]  Yojoa group               6.0 - 13.2
                           Valle de Angeles group    4.0 - 13.1
                           Matagalpa formation       3.2 - 10.0

log [K.sub.sat]            Yojoa group             -4.44 - 3.74
(m [s.sup.-1])             Valle de Angeles group  -4.76 - 4.20
                           Matagalpa formation     -5.22 - 4.17

D (m)                      Yojoa group              1.50 - 4.50
                           Valle de Angeles group   0.50 - 1.50
                           Matagalpa formation      1.50 - 5.00

Parameter                  Rock group              Standard deviation

[[rho].sub.sat]            Yojoa group                   0.027
(Mg [m.sup.-3])            Valle de Angeles group        0.044
                           Matagalpa formation           0.066

[phi](*)                   Yojoa group                    3.20
                           Valle de Angeles group         0.97
                           Matagalpa formation            2.94

[[[tau].sub.sat].sup.kPa]  Yojoa group                    3.5
                           Valle de Angeles group         4.4
                           Matagalpa formation            2.4

log [K.sub.sat]            Yojoa group                    0.34
(m [s.sup.-1])             Valle de Angeles group         0.27
                           Matagalpa formation            0.39

D (m)                      Yojoa group                    n/a
                           Valle de Angeles group         n/a
                           Matagalpa formation            n/a

                            Area       No. landslide  Slide density
                        ([km.sup.-2])     systems     ([km.sup.-2])

Yojoa group                 29.03           43            1.48
Valle de Angeles group       8.06            8            0.99
Matagalpa formation          9.00           30            3.33

Table 2

Surface lithology and [alpha]*W classification

                  Group area    Area prediceted     [alpha]*W
Rock group        ([A.sub.G]  unstable ([U.sub.G]+  class ++

                  [km.sup.2]       [km.sup.2]
Yojoa                29.0             2.7               A
Valle de Angeles     8.1              2.0               A
Matagalpa            9.0              3.2               A
Area totals          46.1             7.9               A

                  Percent of  Percent of       Number of
Rock group        [A.sub.G]   [U.sub.G]   observed slides [n]

                      %           %
Yojoa                76.2        29.8             11
                     12.2        16.2              4
                     11.5        53.5             20
                      0.1         0.5              8
Valle de Angeles     15.9         6.5              0
                     79.4        74.7              5
                      2.5        10.0              2
                      2.2         8.7              1
Matagalpa            45.2        24.8              2
                     45.0        47.4              9
                      9.2        26.1             10
                      0.6         1.7              9
Area totals          61.6        12.4             13
                     28.2        42.4             18
                      9.7        42.3             32
                      0.5         2.9             18

                  Number of observed
Rock group        predicted unstable

Yojoa                     3
Valle de Angeles          0
Matagalpa                 0
Area totals               3

+ Prediceted probability of failure > 0.5

++ Class A: [alpha] < 30 [degrees], W < 0.5; Class
B: a < 30[degrees], W < 0.5; Class C: a < 30[degrees], W < 0.5: Class
D: a < 30[degrees], W > 0.5.

[n] To account for spatial uncertainty in the landslide inventory,
slope-wetness classification are based on the greatest slope and wetness
values found within 35 m of the mapped failure. All slides were mapped
after Hurricane Mitch.

Table 3

Preferred management techniques for [alpha] *W classes.

Class  Slope           Wetness            Identifiers

  A    < 30[degrees]   dry to    Coarse or well-drained soil;
                        moist    rolling terrain; dries out
                                 quickly after rain events

  B                    wet or    Clay-rich, cohesive soil:
                      saturated  frequently wet; saturation
                                 and ponding during and after
                                 rainstorms; wetland
                                 vegetation; concave slopes
                                 with large contributing areas.

  C    > 30[degrees]   dry to    Coarse or well-drained soil;
                        moist    steep slopes; dries out
                                 quickly after rain events;
                                 slopes convex, flat, or mildly

  D                    wet or    Clay-rich, cohesive soil;
                      saturated  obstructed drainage; steep,
                                 concave slopes; overland
                                 flow during rainstorms
                                 and ponding in local

Class  Slope           Wetness   Potential for slope failure

  A    < 30[degrees]   dry to    low, unless degraded or
                        moist    disturbed

  B                    wet or    moderate; higher for steeper
                      saturated  slopes within the class and
                                 when the toe slope is
                                 excavated or naturally

  C    > 30[degrees]   dry to    moderate; high on cleared or
                        moist    extremely steep
                                 (> 50[degrees]) slopes.

  D                    wet or    high under all conditions

Class  Slope           Wetness     Suggested remediation measures

  A    < 30[degrees]   dry to    Generally none necessary. Good
                        moist    locations for infrastructure or
                                 intensive agriculture. Surface
                                 erosion is the primary soil
                                 management concern.

  B                    wet or    Gently sloped grass or stone lined
                      saturated  drainage ditches, thick ground
                                 cover with tall grass or mulch.
                                 Trees only valuable if roots reach
                                 to bedrock, and not recommended in
                                 the absence of additional ground
                                 cover. Ensure that roads are well-
                                 drained through outsloping, rolling
                                 dips or ditch/culvert system. Avoid
                                 steep or tall road cuts.

  C    > 30[degrees]   dry to    Incorporation of trees and shrubs
                        moist    into the agricultural system.
                                 Buffer strips of grasses and woody
                                 plants at the bottom slope and
                                 along roads or rivers. Erosion
                                 control netting (if available) for
                                 use in establishing woody
                                 plantings. Re-route roads where
                                 possible, or reinforce the road cut
                                 with vegetation, gabion, or other
                                 engineered measures. Careful
                                 disposal or removal of spoils from
                                 road construction.

  D                    wet or    Candidate areas for protection.
                      saturated  Avoid slope cutting for roads or
                                 terraces, and avoid as sites for
                                 infrastructure, either on or below.
                                 Stone-lined drainage ditches and
                                 thick ground cover are recommended;
                                 trees only beneficial if roots
                                 reach to bedrock. Grass rows should
                                 be oriented diagonally or down
                                 slope to maximize shedding effect;
                                 avoid contour planting, especially
                                 on active slide zones.


The authors thank the Cornell Institute for International Food and Agriculture Development (CIIFAD) for sponsoring this research and thank Teratech Consulting Ltd. (Salmon Arm, British Columbia, Canada) for making the SINMAP model available.

References Cited

Alexander D.E. 1995. A survey of the field of natural hazards and disaster studies. Pp 1-19. In: A. Carrara and F. Guzzetti (ed.). Geographical Information Systems in assessing natural hazards. Kluwer Acad. Pub., The Netherlands.

Anderson, M.J., M.J. Kemp, and D.M. Lloyd. 1988. Applications of soil water finite difference models to slope stability problems. Pp 525-530. Proceedings of the 5th International Symposium on Landslides, Lausanne, Switzerland. A.A. Balkema, Rotterdam.

Bergin, D.O., M.O. Kimberley, and M. Marden. 1995. Protective value of regenerating tea tree stands on erosion-prone hill country, East Coast, North Island, New Zealand. New Zealand Journal of Forestry Science 25(l):3-19.

Carrara, A., M. Cardinali, F. Guzzetti, and P. Reichenbach. 1995. GIS technology in mapping landslide hazard. Pp 135-175. In: A. Carrara and F. Guzzetti (ed.) Geographical Information Systems in assessing natural hazards. Kluwer Acad. Pub., The Netherlands.

Chung, C.F., A.G. Fabbri, and C.J. van Westen. 1995. Multivariate regression analysis for landslide hazard zonation. Pp 107-133. In: A. Carrara and F. Guzzetti (ed.) Geographical Information Systems in assessing natural hazards. Kluwer Acad. Pub., The Netherlands.

COHDEFOR. 1998. Programa de manejo de los recursos naturals renovables de la cornea del embalse "El Cajon": diagnostico general U.O.C. La Libertad.

Collison, A.J.C., M.G. Anderson, and D.M. Lloyd. 1995. Impact of vegetation on slope stability in a humid tropical environment: A modeling approach. Institution of Civil Engineers, Water, Maritime, and Energy 112:168-175.

Cruz, R. and I. Reyes. 2000. Reporte: Efectos del Huracan Mitch a la Cuenca Humuya area de La Libertad (Honduras). U.S. Department of Agriculture Foreign Agricultural Service.

Divakarla, P.K. and E.J. Macari. 1998. GIS based assessment of rain induced landslides. Pp 685-695. In: First International Conference on New Information Technologies for Decision Making in Civil Engineering, Montreal, Canada. October 11-13, 1998.

Duncan, J.M. 1996. State of the art: Limit equilibrium and finite-element analysis of slopes. Journal of Geotechnical Engineering 122:577-596.

El-Ashry, M.T. 1988. Forward. Pp ix-xi. In: W.C. Moldenhauer and N.W. Hudson (ed.). Conservation Farming on Steep Lands. Soil and Water Conservation Society, Ankeny, Iowa.

El-Swaify, S.A. and J.H. Fownes. 1992. Erosion processes and models: applications in the tropics. Pp 135-150. In: H. Humi and K. Tato (ed.). Erosion, conservation, and small-scale farming. Geographica Bernensia, Marceline, Missouri.

Fredlund, D.G. 2000. The 1999 R.M. Hardy lecture: The implementation of unsaturated soil mechanics into geotechnical engineering. Canadian Geotechnical Journal 37:963-986.

Hammond, C., D. Hall, S. Miller, and P. Swetik. 1992. Level I stability analysis (LISA) documentation for version 2.0. Gen Tech. Rep. INT-285, 190 pp. Intermountain Res. Stn., USDA Forestry Service, Ogden, Utah.

Hansen, A. 1984. Landslide hazard. Pp 523-602. In: D. Brundsden and D.B. Prior (ed.). Slope instability. Wiley, New York.

Hellin, J., and M.J. Haigh. 1999. Rainfall in Honduras during Hurricane Mitch. Weather 54(11):350-358.

Heuvelink, G.B.M. and E.J. Pebesma. 1999. Spatial aggregation and soil process modelling. Geoderma 89:47-65.

Holtz, R.D., and W.D. Kovacs. 1981. An Introduction to Geotechnical Engineering. Prentice-Hall, Englewood Cliffs, New Jersey.

Howell, J.H., J.E. Clark, C.J. Lawrence, and I. Sunwar. 1991. Vegetation structures for stabilising highway slopes. Eastern Region Interim Project, Department of Roads, Kathmandu, Nepal.

Iida, T. 1999. A stochastic hydro-geomorphological model for shallow landsliding due to rainstorm. Catena 34:293-313.

IADB (Inter-American Development Bank). 1999. Reducing vulnerability to natural hazards: lessons learned from Hurricane Mitch. Paper presented to: Consultative Group for the Reconstruction and Transformation of Central America. Stockholm, Sweden.

Iverson, R.M. 2000. Landslide triggering by rain infiltration. Water Resources Research 36:1897-1910.

Jibson, R.W. 1993. Predicting earthquake-induced landslide displacements using Newmark's sliding block analysis. Transportation Res. Rec. 1411:9-17.

Keller, G. and Sherar, J. 2000. Manual de Practicas Mejoradas de Caminos Forestales (PMCs). Proyecto de Desarrollo Forestal. ESNACIFOR/USAID, Siquatepeque, Honduras.

Lal, R. 1999. Erosion Impact on Soil Quality in the Tropics. Pp 285-305. In: R. Lal (ed.). Soil quality and soil erosion. CRC Press, Boca Raton, Florida.

Lal, R. 1994. Global overview of soil erosion. Pp 39-51. In: Soil and water science: key to understanding our global environment. SSSA Special Publication No. 41. Soil Science Society of America, Madison, Wisconsin.

Mangurian, D. 1997. Ingenuity saves dam. IDB America Online:

McCown, S., A. Graumann, and T. Ross. 1998. Mitch: the deadliest Atlantic hurricane.

Miles, S.B., D.K. Keefer, and T.L. Nyerges. 2000. A case study in GIS-based environmental model validation using earthquake-induced landslide hazard. Pp 481-492. In: 4th International Symposium on Spatial Accuracy Assessment in Natural Resources and Environmental Sciences. Proceedings Accuracy, Amsterdam, The Netherlands.

Montgomery, D.R. and W.E. Dietrich. 1994. A physically based model for the topographic control on shallow land-sliding. Water Resources Research 30(4): 1153-1171.

Montgomery, D.R., K.M. Schmidt, H.M. Greenberg, and W.E. Dietrich. 2000. Forest clearing and regional land-sliding. Geology 28:311-314.

Nagle, G.N., T.J. Fahey, and J.P. Lassoie. 1999. Management of sedimentation in tropical watersheds. Environmental Management 23(4):441-452.

Pack, R.T., D.G. Tarboton, and C.N. Goodwin. 1998. Terrain stability mapping with SINMAP, technical description and users guide for version 1.00. Report No. 4114-0, Terratech Consulting Ltd., Salmon Arm, B.C. Canada.

Pla, I. 1997. A soil water balance model for monitoring soil erosion processes and effects on steep lands in the tropics. Soil Technology 11:17-30.

Reynolds, W.D. 1993. Saturated hydraulic conductivity: laboratory measurement. In: M.R. Carter (ed.). Soil Sampling and Methods of Analysis. Canadian Society of Soil Science.

Royster, D.L. 1979. Landslide remedial measures. Bull. Association of Engineering Geologists 16:301-352.

Shaxson, T.F. 1988. Conserving soil by stealth. Pp 9-17. In: W.C. Moldenhauer and N.W. Hudson (ed.). Conservation farming on steep lands. Soil and Water Conservation Society, Ankeny, Iowa.

Sheng, T.C. 1989. Soil Conservation for small farmers in the humid tropics. FAO Soils Bulletin 60. FAO, Rome.

Sidle, R.C., A.J. Pearce, and C.L. O'Loughlin. 1985. Hillslope stability and land use. Water Resour. Mongr. Ser., vol. 11. 140 pp. AGU, Washington, DC.

Simon, A., M.C. Larson, and C.R. Hupp. 1990. The role of soil processes in determining mechanisms of slope failure and hillslope development in a humid-tropical forest, eastern Puerto Rico. Geomorphology 3:263-286.

Terlien, M.T.J., C.J. van Westen, and T.W.J. van Asch. 1995. Deterministic modeling in GIS-based landslide hazard assessment. Pp 57-77. In: A. Carrara and F. Guzzetti (editors). Geographical Information Systems in assessing natural hazards. Kluwer Academic Publishers, The Netherlands.

Triomphe, B.L. 1996. Seasonal nitrogen dynamics and long-term changes in soil properties under the mucuna/maize cropping system on the hillsides of Northern Honduras. Ph.D. dissertation. Cornell University, Ithaca, New York.

World Neighbors. 2000. Lessons from the field--reasons for resiliency: Toward a sustainable recovery after Hurricane Mitch. World Neighbors Central America Office, Tegucigalpa, Honduras. 36 pp.

Wu, W. and R.C. Sidle. 1995. A distributed slope stability model for steep forested basins. Water Resources Research 31(8):2097-2110.

Zaitchik, B.F., H.M. van Es, and P.J. Sullivan. 2002. Variability and scale in the application of a slope stability model no Honduras. Soil Science Society of America Journal (In press).

Zhang, W. and D.R. Montgomery. 1994. Digital elevation model grid size, landscape representation, and hydrologic simulations. Water Resources Research 30(4):1019-1028.

Benjamin F. Zaitchik is a graduate student in the Department of Geology and Geophysics at Yale university, New Haven, Connecticut, and Harold M. van Es is a professor in the Department of Crop and Soil Science at Cornell university, Ithaca, New York.
COPYRIGHT 2003 Soil & Water Conservation Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2003 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:geographic information system
Author:Zaitchik, B.F.; Es, H.M. van
Publication:Journal of Soil and Water Conservation
Geographic Code:2HOND
Date:Jan 1, 2003
Previous Article:Earthworm (Lumbricidae) survey of North Dakota fields placed in the U.S. Conservation Reserve Program. (Research).
Next Article:Examining the targeting of conservation tillage practices to steep vs. flat landscapes in the Minnesota River Basin. (Research).

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |