Closer to Home

Having been talking a lot at a huge, global scale, it’s time to turn things on their head and focus in on the national scale. Dear old blighty (the UK, for the uninitiated) naturally has large banks of hydrological research, given the strength of our academia. The question is, what is going to happen to river flow, runoff and flooding in the future? Will we still be able to sit near the local river in driving rain on a ‘summers’ day, eat poor quality ice cream and have a family argument?

In the early 2000s, Defra advised that research in the UK should take account of possible climate change by modelling a 20% increase in peak flows in the next 50 years. Back in 2004, Reynard, et al. (2004) carried out hydrological modelling using UKCIP02 climate scenarios (UKCIP09 are now the most recent) and concluded that all but the most extreme increases in flow across major catchments in the UK would be within this 20% boundary. This would suggest the 20% peak flow increase is a useful precautionary value in the face of great uncertainty of the effect of climate change on river flow. Kay, et al. (2009) identify the variety of sources of uncertainty in hydrological prediction: future greenhouse gas emissions; Global Climate Model (GCM) structure; downscaling from global to regional model level; the structure and parameterisation of hydrological models, and the natural internal variability of climate. Clearly, this is not an easy process with any level of great surety. Both Kay, et al. (2009) and Prudhomme, et al. (2003) concur that the largest uncertainties are attributed to the type of GCM that is used, suggesting predictions for the future should be made using a variety of GCMs under a range of climate scenarios.

With respect to likely effects of climate change on rivers in the UK, Wilby, et al. (2008) suggest a connection between the seasonality of flood and the seasonality of climate in larger catchments; the peak of annual rainfall is followed by annual peak of discharge, as with the River Thames. For smaller catchments, large flow events tend to be more flashy in nature and occur directly after the heaviest rain, whatever time of the year – future changes in precipitation intensity could have implications for flows and flash flooding in such catchments. Prudhomme,et al. (2003) model the largest increases of peak flow in the UK (using a variety of GCMs) to be in northern England and Scotland, with a 0.13% increase in flood magnitude each year, compared to 0.04-0.05% in southern catchments of England. While some of this may be related to catchment morphology, there is a clear geographical signal of climate change across the country. 

After all the global scale examination, we see that even an island as small as the UK will likely be differentially affected by changes in discharge and flood magnitude as a result of climatic change. At least we’ll still have the weather to moan about.

Stressed Out

Both runoff and discharge are important in sustaining water resources around the world; Oki and Kanae (2006) have noted that river discharge is a major part of renewable freshwater resources. As a planet, we are nowhere near using 100% of all the water available – but the problem is one of temporal and spatial issues in the availability of water resources. Such variability includes variations in river discharge over the hydrological year, and downstream water being so polluted that it cannot be reused (in this case, it would be runoff that was the renewable resource, not discharge).

Variability is being exacerbated by increasing variability in precipitation – climate change is leading to more precipitation globally, but distributed unevenly with some areas wetter and some drier (Arnell 1999). However, river runoff will likely decline globally, as much of the land surface will see a reduction in precipitation and more evaporation. Modelling using global climate models and projections of population growth/water use allows tentative predictions of how climatic change may affect the amount of people living under water stress – Arnell’s work suggests the largest stresses will be seen in the Mediterranean, the Middle East, and southern Africa. While this is valuable information, long term predictions (e.g. 2050 and beyond) are difficult as the strength of the climate change signal becomes increasingly less clear with time, and models are very sensitive to water use projections (which are also variable). What is clear is that changes in the amount of runoff draining into a country can have an effect on water resource stress.

However, we have to be careful here. In models predicting future water resource stress, there are many confounding factors – climate change, population growth and water demand. Vorosmarty, et al. (2000) modelled three different scenarios in an attempt to quantify the key factor influencing water resource stress. They did this by inferring water demand from the ratio of water use against discharge. One scenario involved only climate change, another involved only expanded water use (from population growth) with no climate change, and the final involved both climate change and expanded water use (Figure 1). Climate change alone proved to have limited region specific impacts, whereas expanded water use alone increased the demand per unit of river discharge for much of the world. Considering both factors together shows that some interaction between population growth and climate change causes some areas of stress per unit discharge decrease, but the general pattern is still one of pandemic increase. The conclusion is that climate change and resultant changes in discharge and runoff are something of a secondary factor compared to population growth and expanded water use. Climate change and water resource stress is directionless at a global scale – it will both increase and decrease stress depending on the location. 

The relative change in demand per unit of discharge for the three different scenarios. Red indicates an increase in demand, blue a decrease. From Vorosmarty, et al. (2000).

So, an important message; while climate change will have effects on runoff and discharge, many of these effects will be mediated by humans. I personally see global scale models involving both climate change and population trends into the future as limited, because of the huge uncertainties and the large timescales involved. No doubt they are useful for considering the overall direction of change, but there can be considerable regional disagreement of the direction of change between models. This is a conclusion often reached in these posts, but it is more often than not true – climate change and the hydrological cycle are incredibly complicated and uncertain.

Open The Floodgates

We know river discharge is going to change around the globe into the future, as a result of changes in the hydrologic cycle forced by climate change. However a numerical change in discharge means very little to anymore more than a poorly groomed, strangely bearded (see figure), obsessive hydrologist. Potential changes in discharge can affect humans in a number of ways, and one of the more obvious is changes in flood risk. Flooding can often be a bad thing, but in some case it can be important e.g. providing water for agriculture, recharging groundwater supplies.

The mark of many a hydrologist.

Modelling studies have indicated that human-induced greenhouse gas increases have contributed to an increase in the intensity of heavy rainfall across 65% of the Northern Hemisphere, and that in the future many places will be subject to an intensification of heavy rainfall. This is because as greenhouse gas levels rise (of which CO₂) is one, the planet will warm, which means clouds have to have more water in them to be able to rain – so when it does rain, it rains a lot. These changes promote increases in large flood events – more water = higher flood risk. 

Throughout the 20^th century, the frequency of large food events (technically speaking, a flood that would be expected to happen every 100years) have become more frequent, and statistical modelling suggests that is trend will continue to occur as CO₂ levels rise. Already historical changes in radiative forcing may have caused the increase in large food events in the 20^th century. 

Recent work looks at the changes in the magnitude of flood events with a 30 year return periods, and clear patterns emerge (Figure 1). Increases in flooding are projected in central and eastern Siberia, as well as southeast Asia, including India. Decreases in flooding are projected in northern and eastern Europe, and parts of central South America, around the Amazon basin. These patterns are similar to those projected to river discharge – it is no surprise to see the two are interlinked.

Figure 1 - Projected percentage changes in the magnitude of 30 year return flood events. From Dankers, et al. (2014).

These changes in relatively common floods (every 30 years) have to potential to impact humans, whether it is destruction by floodwaters or agricultural drought from a lack of water inundation. We will focus on specific impacts in another post, but it is clear that humans are innately linked to river discharge and flooding.

I leave you with the song that inspired the name of this post.