Temporal analysis for multi-hazard risk assessment of rice cultivation in 1 coastal areas: a case study of Soc Trang, Vietnam 2

8 Rice is an important human crop and rice cultivation is threatened due to natural disasters, 9 leading to negative effects on national and global food security. The natural disasters, such 10 as tropical cyclones and saline intrusion, have dramatic influences in coastal regions. To 11 investigate possible impacts of these disasters on rice cultivation, it needs an efficient tool 12 to assess potential disasters impacts and a risk index is highly applicable. Therefore, this 13 study aims at establishing a risk assessment of rice production in coastal areas under effects 14 of tropical cyclones and saline intrusion. We adopted risk definition introduced by IPCC 15 (2014) in which risk is a function of hazard, exposure and vulnerability. Multiple hazards 16 of tropical cyclones and saline intrusion were indicated by their frequency and severity at 17 some critical levels of 25%, 50% and >50% rice yield reduction. Each hazard was 18 weighted by its damage on rice yield. Exposure and vulnerability of rice crops are 19 evaluated at different growing phases. Tropical cyclone hazard index was ranked high and 20 very high in the wet season while salinity hazard index was ranked very high in the dry 21 season. Due to the combined effects of tropical cyclones and salinity, rice crop is highly 22 susceptible during the reproduction phase and at the panicle initiation stage particularly. 23 Based on the cropping calendar of My Xuyen, the period of October-November was the 24 very high vulnerability period since it had the largest rice cultivable area and rice crops 25 were at the reproduction phase. This result shows that rice crops are at high risk in October 26 and November. Noticeably, saline intrusion reaches the highest level in April and May, but 27 no risk is at this period because of no rice crop cultivated. This can reflect a measure to 28 reduce risk by adjusting the cropping calendar.


11
investigate possible impacts of these disasters on rice cultivation, it needs an efficient tool 12 to assess potential disasters impacts and a risk index is highly applicable. Therefore, this 13 study aims at establishing a risk assessment of rice production in coastal areas under effects 14 of tropical cyclones and saline intrusion. We adopted risk definition introduced by IPCC 15 (2014)

Introduction 31
According to FAO (2017) natural disasters caused damages of agricultural productions, 32 including crops and livestock, about USD 93 billion in developing countries in a decade (2005-33 2014). Among these natural disasters, tropical cyclones are one of the main disastrous 34 phenomena causing considerable damages around the world which cause a damage of USD 26 35 billion annually (Mendelsohn et al. 2012). In addition, there is an annual loss of approximately 36 USD 12 billion due to saline-affected land (Qadir et al. 2008). The mentioned damaged 37 phenomena are the major factors affecting coastal areas. 38 The Vietnamese Mekong Delta (VMD) is located at the downstream of the Mekong River which 39 is one of the largest rivers in the world (MRC 2010). The VMD is a crucial economic region in 40 Vietnam. Moreover, it plays an important role in ensuring food security not only for Vietnam,41 but also for the world since the VMD contributes up to 90% to the annual rice export of Vietnam 42 (GSOVN 2010). With the land resources of approximately 4 million ha, three-quarters of this 43 area is used for agricultural production (Kakonen 2008). Therefore, rice production in the coastal 44 VMD is facing challenges of tropical cyclones and saline intrusion. For example, Soc Trang is a 45 coastal province, dealing with these challenges. Although Soc Trang is a coastal province, rice is 46 the major agricultural products. Consequences, rice cultivation in Soc Trang is at risk of tropical 47 According to the recent reports on natural disaster prevention and control of My Xuyen district, 106 saline intrusion and tropical cyclones are the most dangerous phenomena. These phenomena 107 cause substantial damages to agricultural cultivation recently. Therefore, these two events are 108 considered for multiple hazards assessment. Hazard indexes are calculated by the below 109 equation. 110 (2) 111 where is multiple hazard index; is the hazard index of phenomenon i; is the weight of 112 phenomenon i. The of each phenomenon is computed based on its frequency and intensity. 113 The hazard index of each phenomenon should be normalized due to different range of the hazard 114 index. We used the weights in calculation of hazard index because this help to prioritize the 115 importance of each phenomenon. The weight of a phenomenon is simply identified as damage 116 percentage caused by that phenomenon of the total damage. 117

Tropical cyclones 118
Tropical cyclones are of the most destructive natural hazards not only in Vietnam but also over 119 the world. They affect large areas depending on their intensity and track of the tropical cyclones. (3) 131 where is the number of points of 6-hour tropical cyclones in the circle; is the maximum 133 wind speed near the tropical cyclone center ; is the weight of the tropical cyclone center 134 which was calculated by its inverse distance to the study area; and is the distance between the 135 tropical cyclone center and the study area centroid (km). 136

Saline intrusion 137
Saline intrusion is the most influential factor that impacts agricultural production in My Xuyen According to Ayers and Westcot (1985), the agricultural crops are not affected by the electrical 147 conductivity (EC) of irrigation water which is lower than 0.7 dS/m. When the EC is higher than 148 0.7 dS/m, the agricultural crops begin to be damaged slightly. The crops are severely affected by 149 the EC of higher than 3.0 dS/m. This is a common classification for salt stress of irrigation water 150 for agriculture and this highly agrees with salinity hazard of irrigation water defined by Bauder capacity to adapt to saline water. For instance, effects of saline water on rice yield were 153 9 evaluated by Ayers and Westcot (1985) and is presented in Error! Reference source not 154 found.. Total dissolved solids (TDS) are the common data of water salinity and they were 155 estimated by using its relationship with EC. The average ratio of total dissolved solids and EC is 156 0.64 (Ali et al. 2012 The salinity data collected were used to analyse the monthly frequency of the selected salinity 159 levels. To assess effects of saline water on rice crops, we defined the Salinity Potential Index 160 (SPI) based on salinity levels and their frequency, presented as follows. 161 where is the number of salinity hazard categories, is the number of salinity occurrence of 163 the category , is the category weights which are equal to 0.25, 0.5, and 1 for slight, moderate 164 and severe categories respectively. 165

Susceptibility of rice crops in the rice-based cropping system 166
Rice is the major agricultural crops in My Xuyen district and the rice-based farming systems are 167 double rice crops and shrimp-rice crops. Rice crops are highly susceptible to the effects of 168 tropical cyclones and salinity in irrigation water. Each type of these hazardous events has a 169 different way to influence the rice crops.  The decreased percentages of rice yields were normalized. Values of the vulnerability index were 203 normalized on a scale from 0 to 1 where vulnerability is lowest or highest respectively. These 204 vulnerability are grouped into five categories, presented in Table 3. 205

Results and discussion 207
Many regions are exposed to a number of natural hazards depending on their spatial 208 13 characteristics. There are several specializing hazards in coastal areas such as tropical cyclones 209 and saline intrusion. It is evident that these natural hazards cause huge damage to agricultural 210 production in the coastal areas. Therefore, the outputs of the hazard and vulnerability assessment 211 assist to deal with risks of these phenomenon. In this section, we present results of hazard, 212 vulnerability and risk assessment of the mentioned hazardous events, influencing the rice crops 213 in My Xuyen district.    very high hazard index. The lowest hazard appears in August, followed by a moderate hazard 261 period. Thus rice is cultivated during this period even in the rice-shrimp cropping system. 262 18 263 Figure 5. Daily maximum salinity at Nga Ba Vam Leo. S1, S2 and S3 are critical levels that 264 reduce rice yields of 25%, 50% and 75% respectively. 265 Table 5 shows temporal variations of vulnerability index of a rice crop which indicates potential 267 effects of tropical cyclones and salinity, damaging at different growth phases. We found that rice 268 crops response to tropical cyclones and salinity differently. Rice is most vulnerable to tropical 269 cyclones at the heading stage and to salinity at the panicle initiation phase. Noticeably, the first 270 half of rice crop is highly vulnerable to salinity while the second half is considerably damaged 271 due to tropical cyclones. The combined vulnerability index is averaged and normalized. It shows 272 that the panicle initiation stage has the highest value of the combined vulnerability index, 273 followed by the heading stage with an index of 0.66. 274 Rice cultivated in My Xuyen are usually short-duration varieties which mature in a range of 105-276 120 days. This is coincided with surveyed data in My Xuyen district recently (Nguyen et al. 277 2019). Table 6 presents cropping calendar or the double rice and rice-shrimp cropping system in 278

Vulnerability 266
My Xuyen. The double rice cropping system includes Summer-Autumn (from June to 279 September) and Winter-Spring (October to January) crops. The rice crop in the rice-shrimp 280 cropping system is cultivated from September to December. The cropping calendar of each 281 system was used for computing vulnerability index which is depicted in Table 6

Multi-hazard risk 292
Based on the hazard and vulnerability identification, monthly risk indexes indicate possibility of 293 rice damage or loss due to effects of tropical cyclones and saline intrusion, are presented in Table  294 7  The nature of risks is considerably useful for assessing potential impacts of tropical cyclones and 302 saltwater intrusion on rice production. It is noticed that the damages of agricultural crops vary 303 with timing of the hazards and crop growth stages. Therefore, to reduce the risks, it needs to 304 reduce hazards or/and vulnerability. First, the hazards can be reduced by improving early

Conclusions 316
We assessed multiple natural hazards of tropical cyclones and saline intrusion. These hazardous 317 events are the major factors, causing significant damages on rice crops in My Xuyen district. To 318 assess risk of the multiple natural hazards, we used a framework of risk on rice crops in a coastal 319 area (My Xuyen, Soc Trang). The study primarily focused on evaluation of potential impacts of 320 the mentioned hazards on rice crops. A temporal analysis was embedded in risk assessment 321 because rice crops vary seasonally. 322 The study have shown that rice crop may face extreme salinity from March to June while it deals 323 with tropical cyclones in October and November. These hazards have various effects on rice 324 crops. Rice crops are considerably susceptible to salinity during the first half of the growing 325 period and to tropical cyclones during the other half. In general, therefore, the results show that 326 rice crops are sharply susceptible in the reproduction phase to the combination of salinity and 327 tropical cyclones. For rice cultivation in My Xuyen, it is considerably vulnerable to these hazards 328 in October and November based on the cropping calendar. We found that the rice crops are at 329 high risk in October and November. Our findings are that saline intrusion reaches the highest 330 level in April and May, but the rice crop was at no risk because of no rice crop cultivated during 331 this period. 332 The multiple hazard assessment provides essential information for the case study in order to 333 create an awareness for natural disaster mitigation. In addition, the integrated result of multiple 334 hazards is much more simply than these of each hazard in providing information to planners and 335 decision-makers because many numbers and scales can be confusing and cumbersome to them. If 336 the results of multi-hazards and risk assessment are used efficiently, it can reduce damage or loss 337 of rice crops/yield. However, this needs an effort of users who convey the information to the 338 natural disaster mitigation procedure. An implication of these findings is to assist decision-339 makers, planners and managers in mitigation and adaptation strategies.

Availability of data and materials 344
The datasets generated and/or analysed during the current study are available from the 345 corresponding author on reasonable request. 346