The end

Time is flying, I still remember the first day I start this blog. I have been enjoying this process, because I learn a lot of totally new things through this blog. It seems I chose a broad topic, natural disasters. Through limited posts, I just picked several items I am interested in to have my exploration.  Too frequent to change another type of natural disasters take me lots of time, but I am really glad I gradually get hold of it. Hope all you read this blog have a nice enjoy.

More about U.S. droughts

In regard to palaeoclimatic data tree rings, there was a drought history with some multidecadal aridity periods of California in the past 500 years and since 1900, there have been at least eight multiyear periods of low precipitation (Larsen et al. 2014). Woodhouse et al. 2009 argued that increasing dryness accompany with increasing temperatures relative to both instrumental records and palaeoclimatic records.

Long period droughts bring a large amount of losses to California. There is an economic impacts assessment of the year 2014 has been taken under the third driest year circumstance to estimate losses from dairies, crop production and livestock by using a set of models (Howitt et al. 2014). The summary table show as follow:

Table1  2014 Drought and California Agriculture Summary 

In the context of worse drought situation, some effective measures should be taken into implement to relief the drought stress. Groundwater is really important in U.S., on which 60% of irrigation relies (Scanlon et al. 2012), however the situation of groundwater seems not very well (Figure1). While increasing water storage through managing extra surface water return to aquifers by up to 3 km3  exhibits hope for dealing with droughts and improving sustainability of groundwater resources in the Central Valley (Scanlon et al. 2012).

Figure1  Measured groundwater level changes from predevelopment (~1950) to 2007 in the high plain aquifer 

In conclusion, recent droughts in California are not unparalleled and they might occur as warm-drought type. Droughts will cause a range of money losses, therefore some effective steps should be exerted to mitigate this worse situation.

11 trillion gallons of water needed in California?!

Yes! 11 trillion gallons of water is required regarding to the calculation from NASA satellite data (NASA news, 16 December 2014). If this number can not give you a specific idea, well, it means around 17 million full Olympic swimming pools. Such large amount required water seems hard to believe after severe floods and mudslides occurred in southern California (BBC news, 13 December 2014). So, what about droughts situation in California or U.S.? Does the drought get more severe? Just like the news what I found on the website that said the drought is the worst has seen in 1200 years in California (USA TODAY, 5 December 2014).

Figure1 the U.S. drought condition (Source: United States Drought Monitor)

After analysis among lots of indices such as SPI (standard precipitation index), NPP (net primary productivity) and NCE (net carbon exchange), Chen et al. 2012 indicates that there was no apparent change of droughts in southern U.S. from 1895 to 2007, although there was an increasing trend in drought intensity in many east regions. This change seems to bring a large loss to this area, as we can see from the resultant obviously drop in NNP which can even reach to 40% during extreme droughts (Chen et al. 2012). There is a wetting trend in most parts of U.S. according to the simulation of soil moisture and runoff except the southwest part in U.S. which has a few declining trend of soil moisture and runoff with increasing drought duration and severity (Andreadis and Lettenmaier 2006).

In regard to this opposite trend compared with the rest regions of the U.S., what might be the underlying reason? Kam et al. 2014 indicates that putting a positive phase of the AMO (Atlantic Multidecadal Oscillation) and a negative phase of the PDO (Pacific Decadal Oscillation) and ENSO (El Niño–Southern Oscillation) together, they would play a worse influence on droughts in the southern U.S. A continuous ridge of high atmospheric pressures generate offshore which would transfer storms track from California to Alaska during a majority of current winters may be the cause of droughts, because California gets most of precipitation in Decembers and Januaries (Dettinger and Cayan 2014). They also mentioned that droughts are going to be more severe and frequent due to climate change and there is a deep relationship between droughts and extreme storms in California. An increasing trend of temperatures results in the decreasing trend of mountain snowpack and earlier melting time of spring snow, which would play an important role in a declining trend of total precipitation (Seager and Vecchi 2010).

To sum up, California seems to be a special region in U.S. with different drought trends which  may be not very significant but still increasing and especially in duration and severity of droughts. In addition, there are several causes behind this trend. The first reason is the influence from the positive phase of the AMO, negative phase of the PDO and ENSO; The second is regarded as anomalous atmospheric pressure which further cause the change of storms; The third is global warming which bring effect on amount of snowpack and melting time.

The trend of droughts

IPCC SREX using the PDSI as an index to analyze the condition of observed global droughts and found there are still large uncertainties in  global scale trends. Dai 2012 argued that the prediction from model has consensus with the change of observed global dryness, which illustrate widespread and severe droughts over many regions in following 30-90 years due to declined precipitation and/or increased evaporation.

Figure1 Trend maps for precipitation and sc_PDSI_pm and time series of percentage dry regions. Long term trends during 1950-2010 in annual mean a. observed precipitation b. calculated sc_PDSI_pm using observation-based forcing c. smoothed time series of the drought regions

From Figure2, we can see that the PDSI_Th has a declining trend since 1970s, however this is not same as the PDSI_PM. As for the global area in drought during 1980-2008, the PDSI_Th has a significant increasing trend, although the trend of the PDSI_PM is not very obvious, it still has an increasing trend which is seven times smaller than the PDSI_Th (Sheffield et al. 2012).

Figure 2 Global average time series of the PDSI and area in drought. a, PDSI_Th (blue line) and PDSI_PM (red line). b, Area in drought (PDSI<-3.0) fro the PDSI_Th (blue line) and PDSI_PM (red line). Th: Thornthwaite algorithms. PM: Penman-Monteith algorithms.

However, Sheffield et al. 2012 argued that the PDSI derives from a simple water-balance model which is popular in large scale drought assessment but bring an overestimated result of global droughts and the trend of droughts has just changed a little over the past 60 years according to more further analysis. From paleoclimatic data, it can be found that recent droughts are not unparalleled because several severe megadroughts can be found in the paleoclimatic record (IPCC SREX).

Even regional droughts may bring some global impact, for example, the largest wheat producer and consumer in the world----China has to import grain due to a persistent drought in the northern growing area and world's largest wheat importer Egypt is having some economic and political problems resulting in dramatic increasing trend of food costs (Sternberg 2011).

Droughts

Floods and droughts seem to be two adverse extremes, so after exploring floods let’s turn to another type of natural disasters, droughts. It seems not hard to understand what the word “drought” means. We can easily get a picture in mind of dry earth with multiple deep cracks, withered crops and so on. IPCC Glossary makes a definition of drought: “A period of abnormally dry weather long enough to cause a serious hydrological imbalance.” However, IPCC SREX indicates that there are numerous definitions about drought which bring some difficulties to the research of drought. Heim Jr., 2002 describes droughts as four main categories: 

• meteorological (climatological) drought is an atmospheric condition with none or less than average precipitation
• agricultural drought makes a severe damage to crops although the deep layer of soil may contain adequate moisture
• hydrological drought refers to a persistent period of time of reducing water supply(surface or subsurface, e.g.: streamflow, lakes, reservoir and groundwater)
• socioeconomic drought relates to imbalance between demand and supply of some economic products affected by other three types of droughts

      

        Figure1 Contraction/Desiccation cracks in dry earth       Figure2 A livestock carcass in Marsabitm in northern Kenya,         (Sonoran desert, Mexico)                                                  which has suffered prolonged drought

Drought indices

Although there are some difficulties to explore the change of droughts, because of no consensus on the definition of drought and lack of observation data of soil moisture, there are some drought indices created to help us learn more about droughts.

Standard Precipitation Index (SPI) which is one of broadly used drought indices and solely based on standardized precipitation. The SPI consists of a transformation of long-term historic precipitation records to normal distribution with zero mean and standard deviation of unit (McKee et al., 1993). The SPI is appropriate for quantification of most categories of droughts, because of the strong relationship between itself and different elements (e.g. stream flow, ground water level, soil moisture content) on different time scale (Lloyd-Hughes and Saunders, 2002). Hayes et al., 1999 indicated advantages and disadvantages of the SPI. Three advantages: 1. Simplicity, based solely on rainfall; 2. Versatile, calculated on multiple timescales; 3. Normal distribution, consistent frequencies of severe and extreme drought classifications at any timescales and locations. Disadvantages: 1. depend on the quality of precipitation data; 2. limited data coverage; 3. timeliness of the preliminary data; 4. be misleading in regions with some seasonal rainfall regime; and other limitations.

Table1 Drought classification by SPI value and corresponding event probabilities

Consecutive Dry Days (CDD) is another widely used index which also only based on rainfalls, referring to maximum consecutive days with limited precipitation during a period of time below a certain threshold, typically 1 mm per day (Frich et al., 2002). The advantages of this indicator is that it can be used for a whole year or any seasons which you would like to choose (can see the example in Figure3).

Figure3 Projected annual and seasonal changes in dryness assessed for 2046-2065

(annual time scale) and 2081-2100(annual time scale and two seasons)

Palmer Drought Severity Index (PDSI) is different with two above indices because it considers not only precipitation, but also evapotranspiration. Dai, 2011 compared four forms of PDSI and  found them having a high correlation with soil moisture observation data in the US and Eurasia, also with yearly streamflow and land water storage respectively in most part of objective regions. It is a common index with some good properties, however there are also some limitations. It should be noted that PDSI is appropriate for central US, which means its less compatible property across different areas and over time (IPCC SREX). Therefore, when PDSI is applied for a certain region, it should be calibrated for local situation.

Here, I just make a introduction of these drought indices which are always seen in drought assessment articles. There are many other indicators, such as Precipitation Potential Evaporation Anomaly (PPEA) and Standardized Precipitation Evapotranspiration Index (SPEI). These indices have their own shortcomings and focus on just one or some aspects of droughts, therefore it is good choice to combine local properties to choose indices which can suit the objective region best. However, even though the projection is still not the completely right answer, because the change of environment. 

Losses from floods

Although the change of floods in global scale is still hard to make a certain conclusion, the increasing trend of socioeconomic losses from floods is in high confidence (IPCC WGIIAR5-Chap3). The risk of floods seems to link with numerous factors such as the exposure of population and assets to flood-prone regions, the development of a certain area and the situation of public infrastructures. The flood itself and some influence factors are difficult to make quantitative measurement, however, losses from floods are more easily calculated in comparison with above items. Kundzewicz et al., 2013 indicates that natural disasters related with extreme weather contribute to increasing economic losses. Figure1 clearly shows the upward trends of total and insured losses from floods during 1980 to 2012.

Figure1 Total and insured losses from flood disasters 1980–2012 (in 2012 values).

Considering the contribution of population growth, IPCC SREX analyzes the average physical exposure to river flooding in 60°N~60°S regions with catchments larger than 1000 km2 (areas limited by models). From Figure2, we can find that Asia is the most severe suffering area and affected population of Asia is distinctly larger than other areas. However, Africa has the most rapid growth rate of influenced people, which grows nearly 3.28 times by 2030 than in 1970. Europe and North America have first and second lowest growth rates of average physical exposures among all these regions, whose growth rates are 0.13 and 0.86 respectively. Moreover, Europe and North America are the only two areas whose growth rates under 1. Developing regions seem to suffer more losses than developed regions.

Figure2 Average physical exposure to floods assuming constant hazard (in thousands of people per year).

Above are all in global scale, IPCC SREX has further study in regional scale, choosing Europe as the research object. Figure3 illustrates that Northern and Southern Central Europe have obvious increasing in both expected affected people and economic damage. However, not all areas in Europe have an upward trend of flood risk, some parts of North Europe expected to suffer less during 2071~2100 than before.

Figure3 Impact of climate change by 2071–2100 on flood risk in Europe.

Models always have some limitations which lead to uncertainty of projection of exposed population towards floods. Hirabayashi et al., 2013 tried to use a number of models to make a projection and put all of them together to make a comparison. Although different models have different growth rates of flood exposure, all of them increase by 2100.

Figure4 Global flood exposure for the 20C 100-year flood, or above, in millions. 

The ensemble means of the historical simulations (thick black line) and the future simulations for each scenario (coloured thick lines). 

Not only some certain regions, but also the globe seem to have much losses in the future. The flood may be hard to control, but we can take measures to reduce damage such as relocating people in severe flood-prone regions, improving monitoring and post-disaster reconstruction systems and raising public awareness. I believe the more we do, the less damage we would suffer.

Floods and Climate Change

Thinking of the disastrous aftermath of floods, I would consider whether the incidence of floods get worse or not, especially under anthropogenic influence environment. Milly et al., 2002 pointed out that there is an increasing frequency of great floods mostly in northern high latitudes, which considered under climate change from anthropogenic radiative effects. They even made a projection of extratropical flood frequency using observed data and several models. Figure1 shows clearly about different start points of trends with statistical significance, and all trends from various models turn to be obvious at a certain year. But because of incomplete consideration of other possible forcing such as various land-use, some errors of models and other factors, the conclusion is tentative.

Figure1 Normalized trend of extratropical flood frequency for observations and for scenario experiments. 

                                    (Normalized trends greater than 1 imply significance with respect to this measure. 

                                     Bold curves: historical records; Light curves: continuing operation)

Hirabayashi et al., 2008 uses MIROC model to estimate the change of floods under global warming situation. There seems to be a link between flood frequency and river annual discharge. Increasing frequency of river floods accompanied by increasing annual discharge, but can not be inferred in reverse.