Servicios Personalizados
Revista
Articulo
texto en 
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por emailIndicadores
-
Citado por SciELO -
Accesos
Links relacionados
-
Similares en
SciELO
Compartir
Permalink
Tecnología y ciencias del agua
versión On-line ISSN 2007-2422
Tecnol. cienc. agua vol.9 no.3 Jiutepec may./jun. 2018 Epub 24-Nov-2020
https://doi.org/10.24850/j-tyca-2018-03-04
Articles
Seasonal flood risk assessment in agricultural areas
1Instituto de Ecología, Pesquerías y Oceanografía del Golfo de México (Instituto Epomex), Universidad Autónoma de Campeche, Campus VI, San Francisco de Campeche, Campeche, México, beaevega@uacam.mx
2Instituto de Ingeniería, Universidad Nacional Autónoma de México, Ciudad de México, México, RDominguezM@iingen.unam.mx
3Instituto de Ecología, Pesquerías y Oceanografía del Golfo de México (Instituto Epomex), Universidad Autónoma de Campeche, Campus VI, San Francisco de Campeche, Campeche, México, gposadav@uacam.mx
The agricultural sector in Mexico is exposed to various hydrometeorological risks, the negative effects of which can reach disastrous proportions, causing significant financial loss to producers and the family economy due to the partial or total loss of the investment and the reduction in income from the sale of crops by farmers. These disasters can also damage the regional and national economy by interrupting the production cycle, reducing income, and creating unemployment and food shortages, among others. This study presents a method that evaluates the risk due to flooding in agricultural areas. The method proposed uses a two-dimensional hydrodynamic model to obtain the parameters of the natural hazard, including the depth, duration, and velocity of the flood. Damage curves were constructed from data obtained by administered surveys to scientific, technical, and farming personnel, supplemented with bibliographic information. These curves considered the various stages of the vegetative crop cycle in such a way that by relating them to the probability estimate of an extreme flow occurring each month of the year, it was possible to calculate the expected damage associated with each magnitude of the extreme flows. The study area corresponds to the Champoton River basin in the state of Campeche, Mexico, where the duration of floods, which are slow, is more important than the velocity of the water.
Keywords River flooding; hydrodynamic modelling; damage functions for crops; agricultural damage; flood seasonality; development stages of crops; flood risk; annually expected damage
El sector agrícola en México está expuesto a diversos riesgos hidrometeorológicos, cuyos efectos negativos pueden alcanzar dimensiones de desastres, ocasionando grandes daños patrimoniales a los productores y a la economía familiar, ya sea por pérdida parcial o total de la inversión y del ingreso esperado. También pueden generar daños a la economía regional y nacional por la interrupción del ciclo productivo, reducción de los ingresos, desempleo y desabasto de alimentos, entre otros. Este trabajo presenta una metodología que permite evaluar el riesgo por inundación en áreas de uso agrícola. El método que se propone utiliza un modelo hidrodinámico bidimensional para obtener los parámetros del peligro natural, como son profundidad, duración y velocidad de la inundación. Se construyeron curvas de daño a partir de datos obtenidos mediante la aplicación de encuestas a personal científico, técnico y agricultores, complementadas con información bibliográfica; dichas curvas consideran las distintas etapas del ciclo vegetativo del cultivo, de manera que al relacionarlas con la estimación de la probabilidad de que una creciente ocurra en cada mes del año es posible calcular la esperanza de daño asociada con cada magnitud de la creciente. La zona de estudio corresponde a la cuenca del río Champotón, en el estado de Campeche, México, en la cual las inundaciones, al ser lentas, se caracterizan porque la duración es de mayor relevancia que la velocidad de la corriente.
Palabras clave inundación fluvial; modelación hidrodinámica; funciones de daño en cultivos; daño agrícola; temporalidad de las inundaciones; etapas de desarrollo en cultivos; riesgo por inundación; daño anual esperado
Introduction
The devastating effects of floods are reflected in both the social and economic aspects of a country or region. In recent years, considerable effort has been made worldwide to identify the causes of floods and their characteristics, the degree of affectation, and the preparation that populations undertake when facing these natural phenomena. Moving beyond structural mitigations, the concept of integrated risk management has been developed (van Westen, 2010). Thus, the activities associated with each stage of flood risk management depend particularly on the results of the economic assessment of the impact of floods (Dutta, Herath, & Musiake, 2003; Li, Wu, Dai, & Xu, 2012).
In the agricultural sector, the standard economic damage due to floods tends to be considerably less than in urban zones for the same level of exposure (Merz, Kreibich, Schwarze, & Thieken, 2010). Therefore, even though several methods have been developed to estimate monetary losses in the agricultural sector, damage assessment in rural zones has been carried out using simple and approximate methods (Förster, Kuhlmann, Lindenschmidt, & Bronstert, 2008; Merz et al., 2010; Brémond, Grelot, & Agenais, 2013). This justifies the development of a method that considers the essential elements involved in the production process and that enables estimating the loss of crops due to floods.
To assess flood damage in the agricultural sector, several methods of varying complexity have been proposed, primarily developed in European countries. The differences between the methodologies are a function of the scale of the analysis, size of the study area, accuracy of the analysis, number of resources required, amount of data needed to apply the available numerical models, and number of components in the agricultural system being considered (Meyer & Messner, 2005; Merz et al., 2010). According to several quantitative studies on the impacts of floods on agriculture (Pivot, Josien, & Martin, 2002; Twining et al., 2007; Posthumus et al., 2009; Chatterton, Viavattene, Morris, Penning-Rowsell, & Tapsell, 2010; Morris & Brewin, 2013), the elements which may be directly damaged by a flood are: crops, perennial plant material, soil, buildings, machinery, livestock, animal products, and stored material. Compared to agricultural zones, damage to infrastructure, such as highways and roads, is rarely mentioned, although it is indicated by Förster et al. (2008); Chatterton et al. (2010), and Morris and Brewin (2013). Most studies that evaluate agricultural loss due to flooding mainly consider damage to crops. Dutta et al. (2003) examined other elements, such as damage to farmhouses and agricultural infrastructure, as well as harvest losses. Meanwhile, Pivot et al. (2002) addressed the effects of flooding on soil characteristics, the potential reduction in soil quality, and the loss of soil structure.
For a methodology to provide realistic damage estimates, it is important to select suitable flood risk parameters. Generic parameters and those which can be obtained through hydraulic models are the most widely used and have the greatest influence on the calculation of direct damage (Brémond et al., 2013). There are currently several numerical models available to characterise floods caused by overflowing rivers, for example: the HECRas(Posada, Veja, Ruiz, Echávez-Aldape, & Martínez, 2011; Martínez, 2011); ISIS, Mike 11, and Mike Flood (Patro, Chatterjee, Mohanty, Singh, & Raghuwanshi, 2009; Kadam & Sen, 2012), which are one-dimensional models (1D). Among the most common two-dimensional models are Telemac2, Mike 21, RisoSurf (Sommer et al., 2009), and TrimR2D. The choice of the numerical model to be used will depend on factors such as the quality of the information available and the scale of the study. The flood parameters that can be considered for the construction of direct damage functions for agriculture are: the seasonality of the floods, water depth, duration, current velocity, deposits, environmental pollution, and water salinity (Brémond et al., 2013).
The relative importance of each flood parameter varies from one region to another and depends on the conditions of the flood and the characteristics of the study area. The flood depth is the most widely used parameter in the construction of direct damage functions. In the case of housing, services, and household infrastructure, the percentage damage is associated with the height the water reaches in these buildings. It is also used in agriculture to evaluate damage to plant material, and in some cases to the soil (Brémond, 2012). The seasonality of floods links different damage coefficients to each season. With regards to duration, the number of days the crop is submerged has been used to evaluate the damage to plant material, however, in some cases, it is not clear if this parameter refers to the number of days the soil took to dry, which depends on soil type. Velocity is a parameter that is rarely used in damage functions in agriculture, unlike in the assessment of domestic buildings. In specific cases, ranges are established such as medium, low, and high to determine the potential damage to the plant (Dutta et al., 2003; Förster et al., 2008; Brémond et al., 2013). The deposition of sediments on flood plains is a parameter that can affect production, with repercussions on livestock feed (USACE, 1985). In addition, salt can have specific impacts on crops and soil. In the case of marine intrusion, the effect reduces production because of the toxicity of the salt in the soil (Roca et al., 2011).
To date, no method has included all the parameters mentioned in one damage estimation model (Vozinaki, Karatzas, Sibetheros, & Varouchakis, 2015), however combinations of the most influential parameters have been used. Förster et al. (2008), and Pistrika (2010) considered seasonality and duration, and Citeau (2003) used immersion time, velocity, and depth. Other more complex combinations have been presented by the USACE (1985), which considered seasonality, water depth, duration, velocity, and sediment deposition. Brémond and Grelot (2012) combined four flood parameters to generate their damage functions: seasonality, water depth, duration, and velocity.
For damage estimation methods, the development of damage functions or damage curves is essential and involves a complex process (Smith 1994; Vozinaki et al., 2015). Two types of curves exist: 1) historical, which are obtained from databases on past flood damage, such as HOWAS 21, developed in Germany, and 2) synthetic damage curves, which are based on theoretical analyses of the damage expected under certain flood conditions (Vosinaky et al., 2015). In the past, various damage functions have been used in the agricultural sector, such as functions that have combined historical and synthetic data (Merz et al., 2010), and curves derived from averaged and normalized historical data from the Japanese Ministry of Construction, which take into account the depth and duration of the floods (Dutta et al., 2003). Pistrika (2010) used empirical functions based on data related to agricultural damage, recorded by the Greek Ministry of Rural Development. Brémond and Grelot (2010) constructed damage functions from data obtained from literature and interviews. Thieken et al. (2009) conducted online surveys to create a database on damage in the rural sector. And impact factors for different types of crops have been generated based on empirical data obtained from interviewing experts (Förster et al., 2008).
The agricultural sector in Mexico is exposed to various hydrometeorological risks, the negative effects of which can reach catastrophic dimensions, resulting in property damage for the producers and the family economy, including the partial or total loss of the investment and expected income. Similarly, they can cause damage to the regional and national economy by interrupting the production cycle, reducing profits, and creating unemployment and food shortages, among other effects.
This study aimed to develop a methodology to assess the risk of floods in agricultural areas. The method was based on previous studies adapted to the conditions of the study area and the information available. The proposed method used a two-dimensional model to obtain risk parameters, including the depth, duration, and velocity of the flood. The damage functions took into account the seasonality of the floods relative to the vegetative crop cycle. Due to the lack of a historical database on agricultural damage, damage curves were constructed through interviews conducted with scientific, technical, and agricultural personnel, supplemented with bibliographic information. The study area corresponded to the Champoton River basin in the state of Campeche, Mexico, where floods are slow; hence the parameter of duration is more important than velocity in the agricultural areas.
Study area
The proposed methodology was applied to the Champoton River basin, located in the central part of the state of Campeche (Figure 1). It belongs to hydrological region (HR) number 31 Yucatan West, of the National Water Commission (CONAGUA, 2015), which has a total surface area of 649 km2, with a maximum elevation of 120 m above sea level. To the north, the area is delimited by HR number 32 Yucatan North, to the south by HR number 30 Grijalva-Usumacinta, and to the east and west by the Gulf of Mexico.
The river is approximately 48 km long, from its origins near the town of San Juan Carpizo to its mouth in the Gulf of Mexico (Posada-Vanegas, Vega-Serratos, & Silva-Casarín, 2013). The river mainly follows a smooth slope, and has a maximum average width of 50 m and a mouth with a maximum width of 80 m; its average depth is 4 m. Along the river, there are several freshwater springs, intermittent currents, the Nayarit de Castellot lagoons, known locally as Nilúm lagoon, and the Noch lagoon, as well as the Xbacab, Chuina, and Hool waterfalls (Ramírez, 2015). Its general trajectory is east-west and it is fed underground by the Desempeño and Las Pozas rivers, which begin in hydrological region number 31 Yucatan West and flow towards the Champoton River. Gleysols soils predominate, with a high clay content and low drainage capacity, which leads to rapid saturation and impermeability. Lictic rendzina is found towards the coast, where there is a thin layer of shallow limestone (Sagarpa, 2009).
Methodology
The method proposed for estimating the risk of flood loss in agricultural zones is divided into three sections: the first part assesses the flood hazard by applying a two-dimensional numerical model, which is based on design hydrographs for various return periods and a digital topo-bathymetric model which integrates the topography of the river basin, the bathymetry of the Champoton River, and the coastal zone near the river’s mouth in the Gulf of Mexico. The second part of the two-dimensional hydrodynamic model consists of the deduction of flood damage functions for the most important crops in the Champoton River basin. For this stage, an analysis of the historical production of all the crops was performed for the study area, and a literature search was carried out on the different methods for assessing agricultural damage and the most important variables considered. Field trips were also made to conduct interviews with farmers, researchers. and state and federal officials whose work was associated with agriculture. Finally, reviewing these elements and the information available, criteria were adapted to construct crop damage curves, in particular, the relationship between the development stage of the crop and the likelihood of significant floods associated with each stage. The third stage consisted of constructing a model to evaluate the annual expected flood damage in agricultural zones, which considers the first two stages, land use, and vegetation, as well as the expected production costs. Figure 2 presents the main steps of the methodology proposed to assess flood damage in agricultural zones.
Flood hazard assessment
With regard to risk, a hazard is defined as the probability of occurrence of a potentially damaging phenomenon of a certain intensity, over a given period and at a particular location. It is important to define the disturbance through quantitative parameters with a precise physical meaning that can be measured numerically and be associated by physical relationships with the effects of the event on exposed goods (Guevara, Quaas, & Fernández, 2006).
To characterise flood risk, this study used the Mike 21 Flow Model FM (DHI, Water & Environment, 2014a), which was developed by the Danish Institute of Hydraulic Engineering. This has a complete modelling system for 2D free surface flows, applicable to the simulation of hydraulic and related phenomena in lakes, estuaries, bays, coastal zones, and seas (DHI, Water & Environment, 2014b). The hydrodynamic module (HD) was used to simulate the variation in water levels and flows subject to a variety of boundary conditions in the coastal area and the river. Figure 3 shows the flow diagram of the applied procedure.
The complete topo-bathymetric model was generated by assembling a Digital Elevation Model (MDE, 20 x 20 m pixel) of the Champoton basin, with the Champoton River bathymetry obtained from field measurements using an echo sounder and GPS, from the coast of the Gulf of Mexico near the river mouth to the coastal zone of the city of Champoton (Posada-Venegas et al., 2013). The bathymetry of this zone was complemented by Nautical Chart 28260, which corresponds to the Gulf of Mexico-Barra Tupilco to Isla Piedra, 1981, from the Mexican Secretary of the Navy (SEMAR) (Figure 4). The study area was discretized into a grid of triangular cells of varying size, which provides greater detail of the zones of interest, such as the river, the city of Champoton, the river mouth, and the coastline. The model was calibrated based on the historic floods caused by Hurricane Isidore (2002), field data, storm surge (Posada-Vanegas et al., 2013), and the measured runoff hydrograph.
Design hydrographs
Frequency analysis is used as a tool to predict the future behavior of the flows generated by hydrometeorological events based on historical information. Statistical procedures are used to calculate the magnitude of the flow associated with a return period (Tr), where Tr is defined as the average number of years it takes for a given event, x, to be equaled or exceeded in magnitude at least once during that time period (Escalante-Sandoval & Reyes-Chávez, 2002). The methodologies described by Domínguez et al. (2008) were used to generate the design hydrographs.
Water Flows
To characterize the flood hazard, records of measured flows from the Canasayab hydrometric station (Canasayab HS) on the Champoton River from the period of 1956-2011 were used. This station is located 39 km from the mouth of the Champoton River. There are no tributaries between the location and the study area. Based on mean daily flow records from the Canasayab HS, a frequency analysis of the maximum annual flows was performed. The AX V.1.05 program (Jiménez, 1992) was used to carry out the analysis of probability, which resulted in the double Gumbel function, providing the best fit to the selected data set, with the lowest standard error. Events for the return periods of 2, 5, 10, 20, 50, 100, 200, and 500 years were subsequently calculated (Figure 5).
The flood simulation process is described in Ramírez (2015). The output parameters used were the maximum water depth, flow duration, and velocity in each of the cells of the flexible mesh.
Agricultural flood damage functions
The development of flood damage functions is one of the most important procedures in modeling to estimate the loss in agriculture (Yu, Qin, & Larsen, 2013). In some developed countries, historical data on flood damage exist for the various systems affected, which allows damage curves to be generated (Meyer, Scheuer, & Haase, 2009; Yu et al., 2013; Vozinaki et al., 2015). For the present study, no database was available on crop damage due to floods, hence the criteria and methodologies were adapted, similar to those of Förster et al. (2008), Brémond et al. (2013), and Chau, Cassells y Holland (2014) for the construction of crop damage curves. Figure 6 shows the proposed sequence for the development of the functions.
Flood hazard parameters that affect crops
The selection of the hazard parameters that most influence direct agricultural damage is crucial in order to produce a realistic estimaion. In this study, the selection criteria for the parameters were based on the information collected in the field from farmers, researchers at the Centro Experimental Edzná-INIFAP (Campeche), and from a literature search (Förster et al., 2008; Brémond et al., 2013; Chau et al., 2014; Vozinaki et al., 2015). In accordance with the information obtained and with the characteristics of the study area, the hydraulic parameters that most affected the agricultural zones were water depth, flood duration, and seasonality (temporality). With respect to the water depth, damage begins with the drowning of the plant’s root. The length of time that the plant is submerged, regardless of the water depth, generates hydric stress. Temporality refers to the crops’ growth stage at the time of the flood. For the study area, due to the flat slope of the flood zone, flow velocity did not significantly affect damage.
Types of crops
As previously mentioned, there is no historical floods database in Mexico that indicates the damage to the agricultural sector as a function of the hydraulic parameters that primarily affect crops.
In order to estimate the expected losses due to flooding in the agricultural zones in the study area, information is needed on the types of crops grown in the arable areas. The decision of farmers as to the type of plants to cultivate in a region depends on the cost-benefit relationship, the specific objectives of the farmer, and the characteristics of the soil (Förster et al., 2008).
In an initial stage, official sources of information from different government entities were consulted. One of the consultation services, provided by the Secretariat of Agriculture, Livestock, Rural Development, Fisheries and Food (SAGARPA), is the Agricultural and Livestock Information System (SIAP, March 2015), which provides historical statistical information about the production sectors in Mexico.
The database from 1980 to 2013 was extracted for all the crops grown in the study area, separated into spring-summer (S-S) and autumn-winter (A-W) cycles and irrigation technique. Figure 7 shows that corn is the most representative grain, followed by sorghum, palay rice, soybeans, beans, and several vegetables in a lower percentage, for seasonal irrigation. Hence, corn was chosen to develop the methodology for the construction of damage curves.
Figure 7 Area planted in the DDR Champoton, 2013. Seasonal irrigation mode. (Source: SIAP-Sagarpa, 2014).
Crop growth stages
The impact of the damage on crops depends on their development or growth stage at the time of the flood. According to the International Corn and Wheat Improvement Center (IMWIC, April 2016), the different growth stages of corn can be divided into two categories: vegetative (V) and reproductive (R). Furthermore, they can be grouped into four main periods:
Growth of seedlings (stages VE and V1)
Vegetative growth (stages V2, V3..., Vn)
Flowering and fertilization (stages VT, R0, and R1)
Grain-filling and maturity (stages R2 to R6)
Figure 8 shows the Ritchie and Hanway (1982) scale, which contains the stages considered in this study. The number of days in each stage depends on the particular characteristics of the area of study in terms of the climate conditions, soil type, temperature, and seed variety, among other aspects. In the Experimental Field of Edzná-Campeche of the National Institute of Forestry, Agriculture and Livestock Research (INIFAP, SAGARPA), various materials (seeds) have been tested in central and northern Campeche, including different varieties such as VS-535 (commercial name) and hybrids such as H-431, to determine the average yield of grain in weather conditions ranging from excellent to adverse.
Seeding and harvest dates
For both agricultural cycles, spring-summer (S-S) and autumn-winter (A-W), the areas sown with the first grains must be under rainy conditions. The seeds are planted between the 1st of June and the 31st of July. When the traditional soil tillage system is used, sowing begins when the rainy season is well under way, with an optimum period between the 15th of June and the 15th of July. With the zero-tillage system, this job can be carried out after the first rains, since the mulch from the vegetal residues left on the soil preserves the mositure longer (Experimental Field of Edzná-Campeche, INIFAP. March 2015). Figure 9 shows the calendar for sowing, harvesting, and marketing for the S-S agricultural cycle in 2015 for the corn, sorghum, rice, and soybean crops.
Damage impact
The impact of the damage depends mainly on the development stage of each crop at the time the flood occurs. The impact values range from 0 to 100%. Based on the sowing, harvesting, and marketing calendar (Sagarpa, 2015), the crop development stages were separated on a monthly basis.
The hydraulic parameters of flood depth, duration, and season were included in the construction of the damage functions. Based on the different flood depths, 4 ranges of flood duration were established. A damage function was constructed for each range according to the studies by Vozinaki et al. (2015), Brémond et al. (2013), and Förster et al. (2008) (Table 1).
Table 1 Format of the survey tables used to determine crop damage from flooding.
| % Impact of damage to the crop from flooding | ||||
|---|---|---|---|---|
| Month | Duration (days) | |||
| 1 a 3 | 4 a 7 | 8 a 11 | > 11 | |
| January | … | … | … | … |
| February | … | … | … | … |
| March | … | … | … | … |
| April | … | … | … | … |
| May | … | … | … | … |
| June | … | … | … | … |
| … | … | … | … | … |
| December | … | … | … | … |
During 2012 and 2013, field trips were carried out to collect information from farmers growing mainly corn, sugar cane, and sorghumIn addition, interviews were conducted with researchers from the National Institute of Forestry, Agriculture and Livestock Research (INIFAP), from the Experimental Field of Edzná, who have studied different crops in the state of Campeche and provided their opinions and experiences on the effects a flood can have on crops such as soybean and corn. The sowing and harvest information was complemented by technological packages for sowing corn in the state of Campeche.
The general considerations for the study area were:
The events considered were slow floods, i.e., flow velocity was slow;
Water drainage as a result of the floods was slow;
A single type of crop was planted in each plot or area of cultivation;
The direct damage to the crops was evaluated;
In the study areas, the slope of the land was small, on the order of 0.1%;
The soil in the study area consisted of mollic and eutric gleysol, i.e., with low permeability;
In addition, in the calendar used by Sagarpa in 2015, for the S-S cycle, the sowing season was established between June and July, hence for this study, the start of the production process was considered to be in May with the preparation of the land. Preparations for the A-W cycle began following the harvest in the S-S cycle.
Land use
To delimit the agricultural zones in the study area, a vector dataset on land use and vegetation was used from the National Institute of Statistics and Geography (INEGI), scale 1: 250 000, series V (2011-2012), updated with LANDSAT 2011 satellite images. From this information, the polygons dedicated to agriculture and the type of vegetation cover were obtained, respectively (Figure 10).
Damage estimate
There are several models that estimate flood damage in the agricultural sector, the differences of which lie in the number of variables they consider and in the existence or availability of measured information for their application. The aim of this paper was to assess the agricultural damage caused by floods based on the impact of the damage and the production costs of the crop in the month the events occurred, as a function of the development stages. The following model was used:
Estimate the probability of a flow Q associated with a return period (Tr) in each of the months of the year. This possibility was considered to be proportional to the mean monthly flows recorded at the Canasayab station. The data generated is the value of a probability P(i) for each month i.
Obtain the crop damage percentage (from 0-1) for each month in the event of a flood caused by a flow associated with a Tr (Q Tr ). This data is extracted from the damage function proposed for the different flood durations. Four flood durations were proposed: 1-3d, 4-7d, 8-11d, and >11d. The data generated are called %D (i, d).
Determine the cost per hectare ($/ha) of the crop for the month according to its development stage in the production process. The values obtained are called CPHC(i).
Estimate the annual damage, DC d , for each flood associated with a Tr and for each duration interval, as the weighted sum of the damage corresponding to the 12 months of the year: Equation (1)
Determine the flooded area as a flow Q associated with a Tr for the different duration intervals. This variable is called ACI(j), and is measured in ha, where j varies as a function of the number of Tr analyzed.
Calculate the damage, D, for each duration interval: Equation (2)
Construct risk curves for each duration interval. Each point of the curve has coordinates (
-
Obtain the annually expected damage, DAE, as the area below the risk curve for each duration interval (Meyer, Priest, & Kuhlicke, 2012): Equation (3)
where ΔP j is the increase in the probability of exceedance between return periods.
Obtain the average total damage based on the sum of each duration interval. This is called DTP.
Construct the risk curve with the total average damage for each Tr (DTP, Tr).
The area below the risk curve gives the average annual expected damage, DAEP.
Results and discussion
Hydrodynamic modelling
The modeling, carried out using a hydrodynamic program, made it possible to evaluate the hazard associated with the Champoton River overflowing in the flood zones. The maximum depth values, velocities, and duration of the floods were obtained from each flexible mesh. The measured hydrograph of Hurricane Isidore (2002) was used to calibrate the model, comparing the modeled results with those obtained in the field by interviews conducted with the residents of the agricultural towns and the populations affected. Figure 11 shows the sites where damage surveys were conducted.
Flood seasonality
To obtain the agricultural damage, the production cost and the impact of the damage on the crops were disaggregated by month, and the historical analysis of the flows from the Cansayab HS (1956-2011) was carried out in a comparable manner. Figure 12 shows the distribution of the average monthly runoff values from the entire record. In accordance with Agraz-Hernández et al. (2015), three seasons were identified in the state of Campeche: dry, rainy, and northerlies. During the dry season (February to April) the magnitude of the peak flows is relatively small, but the impact of the damage is greater than during rainy season. Meanwhile, the rainy season is well under way in the months of May to October, and according to the frequency analysis, the maximum annual flows primarily occur in the months of September and October. It is important to mention that during the “northerlies” season, moisture is accompanied by strong cold winds from the northeast, which are present in the months of November to January, and although the order of magnitude of the flows is small compared to those of the rainy season, due to antecedent soil moisture the vulnerability of the agricultural zones is considerable.
The monthly disaggregation of the probability of occurrence of a flood was considered important since the losses in agriculture greatly depend on the month in which the flood occurs (Förster et al., 2008). Table 2 shows the values of the average monthly runoff and their values relative to the annual total. The extreme floods associated with a maximum peak flow occurred in the months of September and October, and according to the average function of damage to corn (Figure 12), the crops’ most vulnerable stage in the S-S cycle occurred during these months, just before and during harvest.
Table 2 Monthly historical flow, Canasayab Station.
| Month | Jan. | Feb. | March | April | May | June | July | Aug. | Sept. | Oct. | Nov. | Dec. |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Average monthly flow | 142.98 | 121.82 | 145.78 | 137.08 | 133.75 | 242.14 | 331.17 | 466.43 | 1465.52 | 1971.99 | 579.30 | 164.88 |
| P(i) | 0.024 | 0.021 | 0.025 | 0.023 | 0.023 | 0.041 | 0.056 | 0.079 | 0.248 | 0.334 | 0.098 | 0.028 |
Damage functions and production costs
As corn is one of the main crops in the study area, the case of analysis for this work was applied exclusively to those farms where this type of crop was identified. The impact of the damage was divided into 4 categories, according to the duration of the flood in the different stages of crop development. According to the results of the interviews, in the months of May and June, the impact of the damage was small for the ranges of duration of 1 to 3 days and 4 to 7 days, since the affectations only involved a delay in the preparation of the land and the sowing date. Meanwhile, the percentages of damage were greater for all the other ranges from July to August. This coincides with the critical period of development of the crop, which goes from 45 to 75 days of age of the plant, during which the crop develops in height and the flowering is established. Then the grain-filling stage occurs between September and October and harvesting begins, depending on the moisture conditions of the grain. The impact of the damage that occurs during this stage is a complete loss (Figure 13).
Figure 13 present the monthly direct costs of the production process. As the crop develops, several activities are added which increase the costs, including, in particular: soil preparation, seed costs, fertilization, weed control, pest control, and labor costs for the harvest.
Damage estimate
Flood damage in agricultural areas was evaluated with Equations 1 to 3, based on the interaction of flood results meshes, land use and vegetation, damage functions by crop type, and production costs.
The calculation process begins with the DEM mesh. In each hydraulic simulation for a certain return period (Tr), a results grid is obtained. Each cell contains information about the maximum flood stress, its duration, and maximum speed. Subsequently, the use of soil is identified for each cell, as shown in Figure 10. Damage functions are applied in each cell, according to the duration interval and the production costs obtained from the technological package for seasonal corn in the state of Campeche (Figure 13).
Next, the proposed methodology for a flood scenario with a return period (Tr) of 10 years is applied.
Table 3 indicates, for each scenario analyzed in the hydraulic simulation, the size of the crop areas disaggregated by the range of duration of the flood.
Table 3 Flooded crop areas
| Tr (years) | Area (ha) | |||
|---|---|---|---|---|
| Flood duration (days) | ||||
| 1-3 | 4-7 | 8-11 | > to 11 | |
| 2 | 0.1488 | 1.96847 | 2.46554 | 6.4838 |
| 5 | 0.25843 | 2.0113 | 4.08826 | 22.42063 |
| 10 | 0.51453 | 3.15503 | 4.20933 | 37.32747 |
| 20 | 0.65314 | 3.67466 | 5.6057 | 48.87276 |
| 50 | 0.9211 | 9.01778 | 14.08794 | 52.27633 |
| 100 | 1.87926 | 16.56971 | 27.21081 | 79.03159 |
| 200 | 11.5739 | 21.13848 | 39.96751 | 81.53425 |
| 500 | 21.4516 | 48.66245 | 52.79741 | 97.70343 |
The following are obtained in the first stage: the damage D (Q Tr ), considering the probability, P(i), that a flood occurs in the different months of the year; the percentage of the impact of the damage, %D (i,d), of a flood in month i with a certain duration; and the production cost of the CPHC (i) crop in month i of development. These are disaggregated monthly and obtained for each flooded cultivated area, ACI (j), for each scenario, Q(Tr), and for each interval of duration, d(i) (Table 4).
Table 4 Damage evaluation in corn crop areas. Scenario Tr = 10 years.
| Month | Jan | Feb | Mar | Apr | May | June | July | Aug | Sep | Oct | Nov | Dec | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| P(i), probability of occurring in month i | 0.024 | 0.021 | 0.025 | 0.023 | 0.023 | 0.041 | 0.056 | 0.079 | 0.248 | 0.334 | 0.098 | 0.028 | |
| % Damage impact (1-3d) | 5 | 10 | 30 | 30 | 2 | 2 | 5 | 10 | 30 | 30 | 2 | 2 | |
| % Damage impact (4-7d) | 10 | 20 | 50 | 100 | 5 | 5 | 10 | 20 | 50 | 100 | 5 | 5 | |
| % Damage impact (8-11d) | 20 | 40 | 80 | 100 | 10 | 10 | 20 | 40 | 80 | 100 | 10 | 10 | |
| % Damage impact (> 11d) | 40 | 80 | 100 | 100 | 20 | 20 | 40 | 80 | 100 | 100 | 20 | 20 | |
| Production cost ($/ha) | 5,740.00 | 6,596.00 | 7,452.00 | 10,252.00 | 2,050.00 | 5,500.00 | 5,740.00 | 6,596.00 | 7,452.00 | 10,252.00 | 2,050.00 | 5,500.00 | |
| Duration (days) | ACI (ha) | Damage($) | |||||||||||
| 1-3d | 0.5145 | 3.58 | 7.00 | 28.41 | 36.75 | 0.48 | 2.32 | 8.28 | 26.82 | 285.59 | 528.67 | 2.07 | 1.58 |
| 4-7d | 3.1550 | 43.87 | 85.90 | 290.32 | 751.13 | 7.33 | 35.59 | 101.60 | 328.88 | 2,918.62 | 10,805.78 | 31.74 | 24.24 |
| 8-11d | 4.2093 | 117.05 | 229.20 | 619.74 | 1,002.13 | 19.55 | 94.97 | 271.11 | 877.56 | 6,230.27 | 14,416.70 | 84.69 | 64.67 |
| > 11d | 37.3275 | 2,075.99 | 4,065.07 | 6,869.63 | 8,886.71 | 346.77 | 1,684.30 | 4,808.28 | 15,563.95 | 69,060.98 | 127,844.28 | 1,501.95 | 1,146.92 |
The sum of the damage is calculated for each duration interval, disaggregated by month. Based on this, the total values for the scenario Tr = 10 years are obtained, as well as the total average DTP damage for all the time intervals (Table 5).
Table 5 Damage by flood duration interval for corn crops. Scenario Tr = 10 years.
| Exceedance probability (P = 1/Tr) | d = 1-3 days | d = 4-7 days | d = 8-11 days | d > 11 days | Average Total Damage, DTP ($) |
|---|---|---|---|---|---|
| 0.1 | 931.55 | 15 424.99 | 24 027.63 | 243 854.84 | 284 239.01 |
The procedure described above was conducted for floods of different return periods and duration intervals. Table 6 summarizes the results obtained, as well as the average total damage (DTP).
Table 6 Summary of damage by scenario and flood duration interval for corn crops.
| Exceedance probability | Damage ($) | Average Total Damage DTP ($) | |||
|---|---|---|---|---|---|
| P = 1/Tr | d = 1-3 days | d = 4-7 days | d = 8-11 days | d > 11 days | |
| 0.5 | 269.40 | 9 623.88 | 14 073.76 | 42 357.71 | 66 324.74 |
| 0.2 | 467.88 | 9 833.28 | 23 336.54 | 146 470.66 | 180 108.36 |
| 0.1 | 931.55 | 15 424.99 | 24 027.63 | 243 854.84 | 284 239.01 |
| 0.05 | 1 182.50 | 17 965.47 | 31 998.37 | 319 278.51 | 370 424.85 |
| 0.02 | 1 667.63 | 44 088.07 | 80 416.56 | 341 513.53 | 467 685.80 |
| 0.01 | 3 402.36 | 81 009.58 | 155 324.33 | 516 301.68 | 756 037.96 |
| 0.005 | 20 954.31 | 103 346.37 | 228 141.93 | 532 651.19 | 885 093.81 |
| 0.002 | 38 837.69 | 237 911.50 | 301 377.37 | 638 282.05 | 1 216 408.62 |
Figure 14 shows the curves that enable obtaining the risk in terms of the expected annual damage to the corn crop. The risk curve was constructed for the average total damage for each Tr, with coordinates (DTP, Tr). Equation 3 was applied to calculate the value of the expected annual average damage (DAEP), which resulted in $ 102 494.29 / ha.
Conclusions
In the present study, a novel methodology was applied to zones where information on the flooding of crop farms was limited, supported by two-dimensional hydrodynamic simulations. The proposed method enables the annually expected damage to be estimated, taking into account the duration, seasonality, and depth of the flood in order to construct damage functions. Due to the lack of a database on damage in the agricultural sector, criteria cited in the literature and from interviews with farmers, researchers, and officials in the agricultural sector were used as a reference. The construction of damage functions was based on a combination of the development stage of the crop and four flood duration ranges in order to quantify the effects, ranging from a small decrease in production to the total loss of the crop. With respect to the parameter of water depth, in the analysis the damage was considered to occur when the cell was wet to the maximum depth, since the damage caused by flooding starts when the oxygen in the ground decreases, which is generated by deterioration in the root of the plant, and consequently, an inadequate development and yield.
The model proposed for the assessment of crop losses was applied to the Champoton River basin, which due to its location has been exposed to various historical floods associated with the presence of hurricanes. Some of the most severe tropical cyclones recorded are Gilberto (1998), Opal and Roxanne (1995), and Isidore (2002). The results of this study are important for the flood risk assessment of the Champoton River and can be used by decision-makers in different government entities to establish better management plans as well as to adjust the costs of premiums paid for agricultural insurance.
It is important to mention that the availability of information on land use is limited, since there is no detailed typology of the crops in the zone, favoring simplifications in the methodology. However, a more detailed analysis is recommended for the crops that are grown in the region, such as sorghum, rice, soybean, pasture, and sugar cane, ones that consider more complex cropping patterns that include the rotation of crops between the spring-summer (S-S) and autumn-winter (A-W) cycles. Similarly, other elements that intervene in agriculture need to be involved in the damage estimate, for example: agricultural machinery, buildings, and silos, among others. The above implies generating more detailed information in the field at the plot level.
Finally, historical databases on damage in agricultural areas need to be improved, which would allow the results obtained to be validated. Similarly, since they are a fundamental part of estimating damage, the proposed damage functions or curves should be updated as much as possible with databases that are being generated and with experimental data.
Acknowledgements
Thanks to the Programa de Mejoramiento del Profesorado (PROMEP-SEP), for the economic support awarded to Beatriz Edith Vega Serratos to carry out her doctorate. To M.C. Débora Libertad Ramírez Vargas for support with hydraulic modelling. Also thanks to the INIFAP personnel, Edzná-Campeche experimental field. The authors thank the academic peers for the reviews made to this work.
REFERENCES
Agraz-Hernández, C. M., Chan-Keb, C. A., Iriarte-Vivar, S. I., Posada- Vanegas, G., Vega-Serratos, B. E., & Osti-Sáenz, J. (2015). Phenological variation of Rhizophora mangle and groundwater chemistry associated to changes of the precipitation. Hidrobiológica, 25(1), 49-61. [ Links ]
Brémond, P., Grelot, F., & Agenais, A. L. (2013). Review article: Economic evaluation of flood damage to agriculture- review and analysis of existing methods. Nat. Hazards Earth Syst. Sci., 13, 2493-2512. Recuperado de https://www.nat-hazards-earth-syst-sci.net/13/2493/2013/ [ Links ]
Brémond, P., & Grelot, F. (2010). Comparison of a systemic modelling of farm vulnerability and classical methods to appraise flood damage on agricultural activities. In: Advancing Sustainability in a Time of Crisis. 11th biennial conference of the International Society for Ecological Economics, August 22-25, Oldenburg and Bremen, Germany. Recuperado de https://hal.archives-ouvertes.fr/hal-00615462/document [ Links ]
Brémond, P., & Grelot, F. (2012). Taking into account recovery to assess vulnerability: application to farms exposed to flooding. In: R. Seppelt, A. A. Voinov, S. Lange, D. Bankamp, & International Environmental Modelling and Software Society (eds.) (Managing Resources of a Limited Planet). 2012 International Congress on Environmental Modelling and Software, Leipzig, Germany. Recuperado de http://www.iemss.org/sites/iemss2012//proceedings/H3_0429_Bremonda_%20Grelot.pdf [ Links ]
Citeau, J. M. (April, 2003). A new flood control concept in the Oise catchment area: Definition and assessment of flood compatible agricultural activities. FIG Working Week, Paris, France (pp. 13-17). Recuperado de https://www.fig.net/resources/proceedings/fig_proceedings/fig_2003/TS_14/TS14_5_Citeau.pdf [ Links ]
Chatterton, J., Viavattene, C., Morris, J., Penning-Rowsell, E., & Tapsell, S. (2010). The costs of the Summer 2007 Floods in England (Technical report). Environment Agency. Recuperado de http://nationalfloodforum.org.uk/wp-content/uploads/EA-Costs-of-Flooding.pdf [ Links ]
Chau, V. N., Cassells, S., & Holland, J. (September 5, 2014). Economic impact upon agricultural production from extreme flood events in Quang Nam, central Vietnam. Nat Hazards. DOI: 10.1007/s11069-014-1395-x [ Links ]
DHI, Water & Environment. (2014a). Mike 21 y Mike 3 Flow Model FM. Hydrodynamic and Transport Module. Scientific Documentation (10 ed.). Hørsholm, Denmark: DHI, Water & Environment. [ Links ]
DHI, Water & Environment. (2014b). Mike 21 y Mike 3 Flow Model FM. Hydrodynamic Module. Step-by-step training guide. Hørsholm, Denmark: DHI, Water & Environment . [ Links ]
Domínguez, R., Esquivel, G., Méndez, B., Mendoza, A., Arganis, M. L., & Carrizosa, E. (2008). Manual del modelo para pronóstico de escurrimiento. Series. México, DF, México: Instituto de Ingeniería de la Universidad Nacional Autónoma de México. [ Links ]
Dutta, D., Herath, S., & Musiake, K. (2003). A mathematical model for flood loss estimation. Journal of Hydrology, 277(1), 24-49. DOI: 10.1016/s0022-1694(03)00084-2 [ Links ]
Escalante-Sandoval, C. A., & Reyes-Chávez, L. (2002). Técnicas estadísticas en hidrología. México, DF, México: Universidad Nacional Autónoma de México, Facultad de Ingeniería. [ Links ]
Förster, S., Kuhlmann, B., Lindenschmidt, K. E., & Bronstert, A. (2008). Assessing flood risk for a rural detention area. Natural Hazards and Earth System Sciences, 8, 311-322. DOI: 10.5194/nhess-8-311-2008 [ Links ]
Guevara, O., Quaas, W., & Fernández, V. (2006). Lineamientos generales para la elaboración de atlas de riesgos. En: Conceptos básicos sobre peligros, riesgos y representación geográfica. Serie: Atlas Nacional de Riesgos. México, DF, México: Secretaría de Gobernación. [ Links ]
Jiménez, E. M. (1992). AX (Ajuste de Función de Probabilidad) versión 1.05. México, DF, México: Centro Nacional de Prevención de Desastres, Facultad de Ingeniería de la Universidad Nacional Autónoma de México. [ Links ]
Kadam, P., & Sen, D. (2012). Flood inundation simulation in Ajoy River using Mike Flood. ISH Journal of Hydraulic Engineering, 18(2), 129-141. [ Links ]
Li, K., Wu, S., Dai, E., & Xu, Z. (September, 2012). Flood loss analysis and quantitative risk assessment in China. Natural Hazards, 63(2), 737-760. DOI: 10.1007/s11069-012-0180-y [ Links ]
Martínez, A. (2011). Estudio hidrológico e hidráulico de la cuenca del río Champotón: mapas de peligro y sistemas de información geográfica (tesis de licenciatura). Campeche, México: Universidad Autónoma de Campeche, Facultad de Ciencias Químico-Biológicas. [ Links ]
Merz, B., Kreibich, H., Schwarze, R., & Thieken, A. (2010). Review article “Assessment of economic flood damage”. Natural Hazards and Earth System Sciences, 10, 1697-1724. DOI:10.5194/nhess-10-1697-2010 [ Links ]
Meyer, V., & Messner, F. (2005). National Flood Damage Evaluation Methods. A Review of Applied Methods in England, The Netherlands, The Czech Republic and Germany (FLOODsite Project Report Contract No:GOCE-CT-2004-505420). Leipzig, Germany: The Helmholtz-Centre for Environmental Research GmbH-UFZ, Department of Economics. [ Links ]
Meyer, V., Scheuer, S., & Haase, D. (2009). A multicriteria approach for flood risk mapping exemplified at the Mulde River, Germany. Natural Hazards, 48,17-39. [ Links ]
Meyer, V., Priest, S. & Kuhlicke, C. (2012). Economic evaluation of structural and non-structural flood risk management measures: Examples from the Mulde River. Natural Hazards, 62(2), 301-324. DOI:10.1007/s11069-011-9997-z [ Links ]
Morris, J., & Brewin, P. (2013). The impact of seasonal flooding on agriculture: The spring 2012 floods in Somerset, England. Flood Risk Management, DOI: 10.1111/jfr3.12041 [ Links ]
Patro, S., Chatterjee, C., Mohanty, S., Singh, R., & Raghuwanshi, N. S. (2009). Flood inundation modelling using MIKE FLOOD and remote sensing data. Journal of the Indian Society of Remote Sensing, 37(1), 107-118. DOI:10.1007/s12524-009-0002-1 [ Links ]
Pistrika, A. (2010). Flood damage estimation based on flood simulation scenarios and a GIS platform. European Water, 30, 3-11. [ Links ]
Pivot, J.-M., Josien, E., & Martin, P. (2002). Farms adaptation to changes in flood risk: A management approach. Journal of Hydrology, 267(1-2), 12-25. [ Links ]
Posada, G., Veja, B., Ruiz, G., Echávez-Aldape, G., & Martínez, A. (2011). Áreas inundables para la cuenca del río Champotón, Campeche: escenarios actuales y futuros. En: A. V. Botello, S. Villanueva-Fragoso, J. Gutiérrez, & J. L. Rojas-Galaviz (eds.). Vulnerabilidad de las zonas costeras mexicanas ante el cambio climático (2a ed.). México, DF, México: Secretaría de Medio Ambiente y Recursos Naturales- Instituto Nacional de Ecología y Cambio Climático, Universidad Nacional Autónoma de México- Instituto de Ciencias del Mar y Limnología, Universidad Autónoma de Campeche. [ Links ]
Posada-Vanegas, G., Vega-Serratos, B. E., & Silva-Casarín, R. (eds.). (2013). Peligros naturales en el estado de Campeche. Cuantificación y Protección Civil. Campeche, México: Universidad Autónoma de Campeche, Centro Estatal de Emergencias de Campeche-Gobierno del Estado de Campeche, Centro Nacional de Prevención de Desastres. [ Links ]
Posthumus, H., Morris, J., Hess, T. M., Neville, D., Philips, E., & Baylis, A. (2009). Impacts of the summer 2007 floods on agriculture in England. Journal of Flood Risk Management, 2, 182-189. [ Links ]
Ramírez, V. D. L. (2015). Modelación de escenarios en las zonas inundables ante la construcción de un desvío carretero en la ciudad de Champotón, Campeche (Tesis de Maestría). Campeche, México: Universidad Autónoma de Campeche, Facultad de Ciencias Químico-Biológicas. [ Links ]
Ritchie, S., & Hanway, J. (1982). How a corn plant develops (Special Report N° 48). Ames, USA: Iowa State University, Cooperative Extension Service. [ Links ]
Roca, M., Bast, H., Panzeri, M., Hess, T. M., Sayers, P., Flikweert, J., Ogunyoye, F., & Young, R. (2011). Developing the evidence base to describe the flood risk to agricultural land in England and Wales, R&D (Technical Report FD2634/TR). London, UK: Department for Environment, Food and Rural Affairs. [ Links ]
Secretaría de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación, Sagarpa. (2009). PRONAC. "Digitalización del Campo Cañero en México para alcanzar la Agricultura de Precisión de la Caña de Azúcar". Desarrollo de un Modelo Integral de SIG y Edáfica como Fundamento de la Agricultura de Precisión en la Caña de Azúcar en México. Etapa I. San Francisco de Campeche, México: Secretaría de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación. [ Links ]
Secretaría de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación, Sagarpa. (2015). Agenda técnica agrícola. Campeche. Campeche, México: Secretaría de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación. [ Links ]
Servicio de Información Agroalimentaria y Pesquera-Secretaría de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación, SIAP-Sagarpa (2014). SIAP. Recuperado de http://www.sagarpa.gob.mx/quienesomos/datosabiertos/siap/Paginas/default.aspx [ Links ]
Smith, D. I. (1994). Flood damage estimation-a review of urban stage-damage curves and loss functions. Water SA, 20(3), 231-238. [ Links ]
Sommer, T., Karpf, C., Ettrich, N., Haase, D., Weichel, T., Peetz, J.-V., Steckel, B., Eulitz, K., & Ullrich, K. (2009). Coupled modelling of subsurface water flux for an integrated flood risk management. Natural Hazards and Earth System Sciences, 9, 1277-1290. DOI:10.5194/nhess-9-1277-2009 [ Links ]
Thieken, A. H., Ackermann, V., Elmer, F., Kreibich, H., Kuhlmann, B., Kunert, U., Maiwald, H., Merz, B., Müller, M., Piroth, K., Schwarz, J., Schwarze, R., Seifert, I., & Seifert, J. (May 6-8, 2009). Methods for the evaluation of direct and indirect flood losses. In: Proceedings. 4th International Symposium on Flood Defense: Managing Flood Risk, Reliability and Vulnerability, Toronto, Ontario, Canada. [ Links ]
Twining, S., Procter, C., Wilson, L., Frost, A., Phillips, K., Turner, T., Tiffin, D., Emery, J., Walker, O., Welch, S., & Wells, A. (2007). Impacts of 2007 Summer Floods on Agriculture, Tech. Rep. Final. Cambridge, UK: ADAS. [ Links ]
US Army Corps of Engineers, USACE. (1985). AGDAM, Agricultural Flood Damage Analysis, User’s Manual (Provisional), Computer Program Documentation CPD-48. Davis, USA: US Army Corps of Engineers, Institute for Water Resources, Hydrologic Engineering Center (HEC). [ Links ]
Van Westen, C. J. (2010). GIS for the assessment of risk form geomorphological hazards. In: I. Alcántara-Ayala & A. S. Goudie (eds.). Geomorphological Hazards and Disasters Preventions (pp. 205-219). Cambridge, UK: Cambridge University Press. [ Links ]
Vozinaki, A. E., Karatzas, G., Sibetheros, I., & Varouchakis, E. (2015). An agricultural flash flood loss estimation methodology: The case study of the Koiliaris basin (Greece), February 2003 flood. Natural Hazards, 79, 899-920. DOI: 10.1007/s11069-015-1882-8 [ Links ]
Yu, J. J., Qin, X. S., & Larsen, O. (2013). Joint Monte Carlo and possibilistic simulation for flood damage assessment. Stochastic Environmental Research and Risk Assessment, 27(3), 725-735. DOI: 10.1007/s00477-012-0635-4 [ Links ]
Received: December 09, 2016; Accepted: November 03, 2017
Este es un artículo publicado en acceso abierto bajo una licencia
Creative Commons
