Keywords Flood monitoring · Damage assessment · Remote sensing · GIS ·
Agricultural food · Vegetation indices · Regression model
16.1 Introduction
Flood can be defined as an excess of water in a specific area. Flooding events could
spread water over surrounding areas of rivers, lakes, and streams causing potential
damage to lives and properties. Flood events are primarily caused by heavy rainfall
along with snowmelt, dam failure, and coastal surge (Jeyaseelan 2003). Flood is
arguably the most common, devastating, and frequent natural hazard in the globe,
and the destruction from its impacts is getting worse due to recent climate change
(Sanyal and Lu 2004; (Greenough et al. 2001). The severity of damage due to floods
is higher in places with human settlements and agricultural activates (Hirabayashi
et al. 2013). As the agricultural areas are outside of the coverage of conventional
hazard management system, it is the most vulnerable and affected by flooding
(Al-Sabhan et al. 2003). Effective real-time flood monitoring including forecasting,
inundation mapping, and a warning might be helpful to reduce the potential damage
and loss. Geographic Information System (GIS) are usually built for data acquisition,
storage, processing, and analysis to cover a broad range of applications (Al-Sabhan
et al. 2003). With the application of advanced GIS couple with the recent Web
capability, it is now possible to develop a near- to real-time flood monitoring system.
Such a system might be a combination of systems such as spatial and nonspatial data
acquisition, integration under sophisticated hydrological modeling, and delivery of
flood data, reports, and alert systems. There are some Web-based near-real-time
flood monitoring systems, for instance, Dartmouth Flood Observatory (DMO) and
Flood Network (IFnet) which are developed in recent years based on
geoinformation. These systems have widely been used for flood monitoring and
flood-related research.
Besides monitoring flood events, it is also extremely important to accurately assess
damages it causes, especially in the agricultural field. Understanding the impact of
a flood on crop yield is not only important to estimate the direct production damages
but also essential in understanding its impact on the global food market. With the
increasing globalized food market, even local level impact on crop (food) productivity
could bring global crises. Although the occurrences of natural disasters like floods are
unavoidable, however, if the damages can be estimated accurately, alternative plans
could be implemented to overcome the loss and fulfill the food demands.
Field visit and surveying local farmers has been a primary approach for flood
damage assessments. These methods are highly labor-intensive, expensive, and only
practical within small areas. Shortcomings of the field visit approach in flood crop
loss assessments can be addressed by using remote sensing techniques. Utilization of
remotely sensed images in crop loss assessments will be efficient due to their wide
area coverage, accurate geocoding, frequent revisit, rapid data distribution, relatively
low data cost, and strong crop/land discrimination (Smith 1997).
322
R. M. Shrestha and M. S. Rahman
Agricultural food · Vegetation indices · Regression model
16.1 Introduction
Flood can be defined as an excess of water in a specific area. Flooding events could
spread water over surrounding areas of rivers, lakes, and streams causing potential
damage to lives and properties. Flood events are primarily caused by heavy rainfall
along with snowmelt, dam failure, and coastal surge (Jeyaseelan 2003). Flood is
arguably the most common, devastating, and frequent natural hazard in the globe,
and the destruction from its impacts is getting worse due to recent climate change
(Sanyal and Lu 2004; (Greenough et al. 2001). The severity of damage due to floods
is higher in places with human settlements and agricultural activates (Hirabayashi
et al. 2013). As the agricultural areas are outside of the coverage of conventional
hazard management system, it is the most vulnerable and affected by flooding
(Al-Sabhan et al. 2003). Effective real-time flood monitoring including forecasting,
inundation mapping, and a warning might be helpful to reduce the potential damage
and loss. Geographic Information System (GIS) are usually built for data acquisition,
storage, processing, and analysis to cover a broad range of applications (Al-Sabhan
et al. 2003). With the application of advanced GIS couple with the recent Web
capability, it is now possible to develop a near- to real-time flood monitoring system.
Such a system might be a combination of systems such as spatial and nonspatial data
acquisition, integration under sophisticated hydrological modeling, and delivery of
flood data, reports, and alert systems. There are some Web-based near-real-time
flood monitoring systems, for instance, Dartmouth Flood Observatory (DMO) and
Flood Network (IFnet) which are developed in recent years based on
geoinformation. These systems have widely been used for flood monitoring and
flood-related research.
Besides monitoring flood events, it is also extremely important to accurately assess
damages it causes, especially in the agricultural field. Understanding the impact of
a flood on crop yield is not only important to estimate the direct production damages
but also essential in understanding its impact on the global food market. With the
increasing globalized food market, even local level impact on crop (food) productivity
could bring global crises. Although the occurrences of natural disasters like floods are
unavoidable, however, if the damages can be estimated accurately, alternative plans
could be implemented to overcome the loss and fulfill the food demands.
Field visit and surveying local farmers has been a primary approach for flood
damage assessments. These methods are highly labor-intensive, expensive, and only
practical within small areas. Shortcomings of the field visit approach in flood crop
loss assessments can be addressed by using remote sensing techniques. Utilization of
remotely sensed images in crop loss assessments will be efficient due to their wide
area coverage, accurate geocoding, frequent revisit, rapid data distribution, relatively
low data cost, and strong crop/land discrimination (Smith 1997).
322
R. M. Shrestha and M. S. Rahman