System Robustness Analysis in Support of Flood and Drought Risk Management

Mens, M. J. (2015). System robustness analysis in support of flood and drought risk management (Vol. 14). IOS Press.


Flood and drought impacts are increasing
Floods and droughts cause increasingly large impacts on societies worldwide. The probability of these extreme events is also expected to increase due to climate change. Water management primarily tries to protect against floods and droughts, for example by building flood protection infrastructure and reservoirs. Despite structural measures to prevent flooding and water shortage, 100% protection can never be provided. Therefore, over the past decades, water management has shifted to a risk-based approach. This means that policies do not only aim at reducing the probability of occurrence of floods and droughts, but also include actions to limit the consequences of potential flooding or water shortage. Both types of measures may aid to reduce
flood and drought risk to an acceptable level.
Limitations of a risk approach
Even if the risk is reduced to an acceptable level, extremely large impacts are not avoided, as demonstrated by recent floods and droughts events with devastating impact. A risk approach considers ten casualties per year in 100 years equal to 1000 casualties at once during the same period. However, the latter have a much larger societal impact. Large impacts occurring at once are considered unacceptable when it is difficult to recover from them. Hence, not only the risk but also the potential impacts should be reduced to an acceptable level. There is a need for decision support methods that help avoiding unacceptably large impacts from floods and droughts. Another reason why risk may not suffice as decision-criterion is that it is uncertain, under both current and future conditions. Estimating current risk requires assumptions on return periods of events that do not occur in measured data. Furthermore, it is uncertain how risks develop into the future, because of uncertain future climate (and climate variability) and socio-economic developments. It is therefore difficult to decide on the most cost-effective strategy in terms of the effect on risk. This further underpins the need for additional decision criteria that take uncertainty into account.

Robustness: a new perspective on dealing with extreme events
The concept of robustness seems useful for dealing with extreme events. Robustness is known from other areas such as engineering and biology, where networks or systems have to maintain their functionality even when some components fail. Areas prone to floods or droughts can be understood as systems. When these systems can remain functioning during flood and drought events, it is likely that unmanageable impacts (i.e., disasters) are avoided. In this thesis, the concept of robustness is made operational by proposing quantifiable criteria. These criteria were tested in two flood cases and two drought cases. The cases have demonstrated the applicability of the framework and have provided insight into the characteristics that influence system robustness. Furthermore, the case studies demonstrated that assessing system robustness may change the preference ordering of management strategies.
Robustness = resistance + resilience
In this thesis, system robustness is defined as the ability of a system to remain functioning under a large range of disturbance magnitudes. Disturbances in this thesis are flood waves in river valleys that may cause flooding, and droughts (resulting from precipitation deficit or streamflow deficit) that may cause water shortage. ‘To remain functioning’ means either no impact from the disturbance or limited impact and quick recovery. System robustness is a function of two other characteristics: resistance and resilience. Disturbances that cause no impact are in the resistance range; larger disturbances that cause limited impact from which the area can recover are in the resilience range. Robustness analysis aims to identify these ranges for a specific system.
Three criteria to quantify robustness
To obtain insight into robustness, this thesis proposes three criteria to describe a system’s response to disturbances,

1. The resistance threshold is the point where the impact becomes greater than zero;
2. The proportionality refers to the graduality of the response increases with increasing disturbance magnitudes;

3. The manageability is the ability to keep the response below a level from which recovery is difficult or impossible.
The first criterion refers to the smallest disturbance magnitude causing significant impacts and is strongly related to the system’s design standard (e.g., protection against floods or reservoir capacity to prevent water shortage). The second criterion originates from the flood risk literature; sudden floods are considered undesirable because people have too little time to prepare, leading to large impacts. Sudden events should thus be avoided in a robust system.
The third criterion compares the impact with a critical recovery threshold. This threshold represents the physical and socio-economic capacity to recover from the impacts of floods and droughts. When impacts exceed the critical threshold, it is assumed that the recovery time is long and that long-term impacts will be unacceptably high.
A robustness perspective may change decisions 

In flood risk management, measures are often prioritized based on risk (a metric that combines flood probabilities and corresponding impact), in comparison to the investment costs. Both flood cases showed that a variety of measures may reduce the risk, but not all of those measures enhance system robustness. This means that different measures may be preferred when their effect on system robustness is also taken into account.
In drought risk management, measures are often assessed on the resulting water supply reliability (i.e., the probability of meeting water demand). The drought cases have demonstrated that not all measures that increase the supply reliability also reduce the drought impacts over the full range of plausible drought events. Thus, different measures may be preferred when their effect on system robustness is also taken into account.
What characterizes a robust flood risk system?
Systems with high protection levels for the entire river valley have high resistance against flood waves. However, when protection levels are equal everywhere, sudden floods can still occur and affect a large and/or vulnerable area. Such a system is not considered robust to flood waves. Robustness of a system with a high resistance threshold can be increased by differentiating protection levels, so that least-vulnerable areas will flood first and more-vulnerable areas are relieved. Another option is to build virtually unbreachable embankments. This prevents sudden flooding and limits the inundation and thus the impact. A combination of unbreachable embankments that are also differentiated in height will further increase robustness to extreme floods. Finally,
measures aimed at impact reduction increase robustness when they reduce the impacts below the recovery threshold.
What characterizes a robust drought risk system?
Drought risk systems have a high resistance threshold when their storage capacity is compared to the demand, for example systems with large reservoirs. The resistance threshold is related to the supply reliability. A variety of supply sources will increase the supply reliability and the resistance threshold. When the objective is to reduce impacts from extreme drought events, demand reduction and temporary measures are more effective than increasing supply on a structural basis. In agricultural drought risk systems, crop diversity and having alternative sources of supply will
enhance robustness to drought.
In conclusion, this thesis contributed to decision making in flood and drought risk management, by developing and testing an additional decision criterion. A robustness analysis method supports the assessment of impacts from extreme events, and is on flood and drought risk systems. A robustness perspective supports makers in exploring low-probability/high-impact events and considering these impacts are societally acceptable. Quantifying robustness inspires the of strategies that reduce flood and drought risk in a way that disasters avoided.