Thursday, 17 December 2015

Spoke Too Soon? Climate Change Impacts on Groundwater Contamination

As discussed in my previous post, with a warming atmosphere, very intense rainfall events (i.e. those in the upper quantiles of the current rainfall distribution) are projected to increase (Allan and Soden, 2008; Allan et al., 2010). This increase is expected to be greatest in the tropics where the warmer starting temperatures will lead to relatively greater rises in the water-holding capacity of the atmosphere (saturation vapour pressure) and thus in the moisture content of the atmosphere, as stipulated by the Clausius-Clapeyron relation (Taylor et al., 2009).

Currently, groundwater commonly has better microbiological quality (lower pathogen content) than surface water, and is therefore more suitable for human consumption (Flynn et al., 2012). However, there is evidence that the projected increase in intense rainfall under climate change will negatively impact on the microbiological water quality of groundwater. To demonstrate this, I shall focus on Bwaise, a typical densely populated informal settlement in Kampala, Uganda (Flynn et al., 2012). Over 200 protected springs supply (partially or fully) domestic water to 60% of the low-income population of the rapidly urbanising city in sub-Saharan Africa (Howard et al., 2003). The protected springs are the preferred source of domestic water for the local population of Bwaise due to its free cost, shorter distance from dwellings, and greater annual consistency in discharge compared to the piped water supply (Kiyimba, 2006; Flynn et al., 2012).

Sporadic contamination of Groundwater in Bwaise Linked to High-intensity Rainfall

Flynn et al., (2012) undertook a 2-month high-frequency discharge and water quality (chemical and microbiological) monitoring program of groundwater at Bwaise III spring, whilst noting the surrounding rainfall events. Thermotolerant coliform bacteria (TTC) are standard bacterial indicators of faecal pollution. Figure 1a shows TTC counts in the spring’s water samples remained very low in dry periods, with sharp rises occurring during and shortly after the onset of sporadic intense rainfall. This is supported by the findings of Kulabako et al., (2007), who reported intermittent but widespread contamination of groundwater in Bwaise III from faecal sources indicated by TTCs and faecal streptococci (FS), which was especially intense during the rainy season.



Figure 1. Time series plot of themotolerant coliform bacterial (TTC) counts with daily rainfall for Bwaise III Spring, Kampala in 2004, taken from a) Flynn et al., 2012, and b) Taylor et al., 2009.

Taylor et al., (2009) found very similar results during their high-frequency monitoring of the Bwaise III protected spring. They reported ephemeral gross contamination by TTC (>103 colony forming units per 100 ml) following less than 1 hour after rainfall events of >5 mm day-1 (Figure 1b). Particularly intense contamination (>104 colony forming units per 100 ml) was observed after rainfall events exceeding 20 mm day-1. This is true for both sets of results (Figure 1a and 1b) which show particularly large increases during the heaviest rainfall events of 2nd August 2004 and 15th August 2004, when concentrations in spring water rose by at least 3 order of magnitude in less than 3 hours (Flynn et al., 2012; Taylor et al., 2009). Thus, the greater the intensity of the rainfall event, the greater the intensity of contamination, and the higher the risk of water-borne disease contraction.

As such, there is strong evidence for contamination of the groundwater of Bwaise III with TTCs during intense rainfall events, with increasing contamination occurring with increasingly heavy rainfall. Howard et al., (2003) demonstrate that high sanitary risk scores correlate significantly (99% and 95% confidence intervals) to the observed bacteriological contamination during the rainy season for the springs in Bwaise. Additionally, Tumwine et al., (2002) observed the incidence of diarrhoeal diseases to increases substantially during the rainy season in Kampala, with infections with the waterborne pathogen Enterocytozoon bienusi during double that recorded during dry seasons. Tumwine et al. (2005) show that E. bienusi and Cryptosporidium spp. are primarily responsible for persistent diarrhoea in children with AIDS. Thus contamination of groundwater supplies with heavy rainfall can give rise to water borne disease outbreaks that contribute to the 3.1% of annual deaths worldwide arising from unsafe drinking water, poor sanitation and inadequate hygiene (Ashbolt, 2004).

The Mechanism for Contamination

The TTCs, indicative of faecal bacteria, are argued to derive from inadequately contained faeces in Bwaise proximate to the protected spring (Taylor et al., 2009). This is supported by the observation of constantly high concentrations of chloride and nitrate in the spring’s discharge samples (Figure 2). These concentrations are substantially greater than those in local rainfall reported by Kulabako et al. (2007), even when accounting for the concentrating effect of soil-zone evapotranspiration (ET) on rain-fed recharge, and thus point to an alternative source of water recharging the aquifer. The most obvious source within the Bwaise spring catchment is latrine effluent, with the persistently high mineral concentrations reflecting the intense loading of sewage in the spring’s catchment (Flynn et al., 2012).

Taylor et al. (2009) argue that the faecal bacteria are transported to the spring in rapid surface and near-surface flows occurring via preferential pathways. The low infiltration capacities of soils up-gradient and immediately adjacent to the Bwaise III spring (Miret-Gaspa, 2004, cited in Flynn et al., 2012), coupled with high surface gradient and limited vegetation cover over most of the catchment, result in frequent sheet overland (surface) flow during periods of intense rainfall, which is then able to enter groundwater immediately up-gradient of the spring discharge point (Flynn et al., 2012).

Measurements of TTC levels in samples of overland flow to the spring revealed it to have extremely high concentrations ranging from 3.4 x 105 to 8.4 x 107 cfu/100 ml (Kulabako, 2005). The infiltration of these highly contaminated flows in the immediate areas around the spring cause a deterioration in water quality but comprise only a tiny proportion of the total recharge to the spring. This explains the observed consistency in, and thus negligible effects on, the springs discharge and nitrate/chloride concentrations throughout each episode of bacteriological contamination for both Taylor et al., (2009) and Flynn et al., (2009) (Figure 2).



Figure 2. Time series of groundwater discharge, nitrate and chloride levels for Bwaise III Spring, Kampala (Uganda) throughout July and August 2004. (Source)

Increased risk of contamination (and thus risk of diarrhoeal diseases) posed by climate change in Kampala and similar environments

To investigate the specific impacts of climate change on rainfall in Kampala, Taylor et al., (2009) underwent dynamical downscaling of climate projections in Uganda from the HadCM3 GCM using a regional climate model (PRECIS). This showed substantial increases in the frequency of heavy rainfall events producing more than 10 mm day-1 over the 20th century for Kampala, and a decrease in the frequency of medium-rainfall events less than 10 mm day-1 (Figure 2). Furthermore, Mileham et al., (2009) reported that this shift in the distribution of rainfall alone accounts for a 238% increase in the estimated recharge of a catchment in south-western Uganda. Under these projections, an increase in the frequency of contamination of groundwater sources in Bwaise III can be expected.


Figure 3. Frequency distribution of rainfall events for a) historical precipitation (1965-1974) and b) dynamically downscales projections of precipitation from the HadCM3 GCM for the River Mitano Basin, Uganda (Source)

How Increasing Contamination under Climate Change can be Minimised

Despite projections of increasing extreme rainfall, observations in Bwaise III point to several human factors which exacerbate contamination of the spring. Improvement of these factors could potentially minimise the negative impacts of climate change on groundwater quality and increase its potential as an adaptive source of domestic water under climate change.

For example, the area immediate up-gradient of the spring is kept clear of polluting debris and runoff by an elevated spring box. Overland flow is diverted around the spring by the box during periods of intense rainfall, which acts to promote the short-circuiting of water contaminated by waste into preferential pathways (Flynn et al., 2012). Furthermore, the box displays localised evidence of cracking, potentially permitting wastewater surface runoff to infiltrate through cracks in the masonry into the ground immediately up-gradient of the spring (Flynn et al., 2012). Based on this model, repair and maintenance of the protective concrete box surrounding the spring’s headworks would significantly reduce risk of microbiologically contaminated surface runoff impacting on the Bwaise III spring’s water quality.

As a more general point of improvement, the absence of sewerage and wastewater disposal infrastructure in many low-income settlements across Africa means much of the population disposes of sewage through infiltration to the subsurface and thus groundwater (Flynn et al., 2012). 70% of Bwaise III’s population use shared pit latrines, and 10% have no access to sanitation (Kulabako et al., 2010). Furthermore, a large proportion of the pit latrines are of the traditional unimproved type (>80%) which do not meet the basic criteria of hygiene and accessibility for children, disabled and elderly, resulting in informal sewage disposal at the ground surface (Flynn et al., 2012). As contamination is sources from faecal wastewater, improved sanitation and containment of waste could prevent increased contamination of groundwater predicted under climate change. This is supported by Taylor et al., (2009), arguing that during the 2002-2003 cholera epidemic in Uganda, (linked to anomalously intensive rainfall during the short rains associated with the ENSO event of 2002 (Alajo et al., 2006)) the duration and mortality rates of outbreaks were greater in rural areas than in Kampala where there is comparatively better access to medical treatment, sanitation, and safe water. Thus improved community hygiene is critical in preventing increased spread of disease associated with intensifying rainfall.


A final point is that the susceptibility of groundwater to contamination by coliforms in the first place depends strongly on the geological composition and thickness of the overlying aquifer material (Swartz et al., 2003). For example, batch study results in Flynn et al., (2012) confirmed the presence of pure haematite in laterite soils contributes substantially to micro-organism attenuation and inactivation, which may serve to protect underlying groundwater under increasingly intense precipitation. Not only does this suggest that some aquifers may be more resilient to contamination under climate change, but also that depending on the mechanical strength of the soil, laterites with minor amounts of haematite show potential as a substrate for low cost water filtration in African communities (Flynn et al., 2012).

Thursday, 10 December 2015

It's Not All Bad News! The Impact of Climate Change on Groundwater Recharge

For many communities in semi-arid and arid regions of Africa, groundwater is the only reliable source of freshwater for drinking and irrigation purposes where surface water is seasonal or perennially absent (MacDonald et al., 2012).  The long-term sustainability of groundwater resources is dependent on replenishment by recharge, which results from effective precipitation infiltrating the subsurface where hydraulic gradients are downward (Taylor et al., 2013). Currently, natural inter-annual groundwater recharge occurs on a decadal timescale by episodic recharge (Taylor et al., 2013).

Global warming is expected to intensify the global hydrological cycle, through amplification of potential evapotranspiration and intensification of precipitation (Jiménez Cisneros et al., 2014). For the IPCC’s AR4 and AR5 GCM’s, the projected increases in extreme monthly rainfall under climate change are of much greater magnitude than changes projected for mean monthly rainfall (IPCC, 2012; Taylor et al., 2013). This is expected to result in more variable river discharges and soil moisture, exacerbating intra-annual freshwater variability and shortages (Owor et al., 2009) and threatening food security through reduced crop yields (Challinor et al., 2007).

A number of key studies provide evidence for groundwater recharge to be biased towards intense, heavy rainfall. Is it possible that under this climate change scenario, groundwater could offer a source of resilience for freshwater resources?

Owor et al. (2009) correlated decadal datasets of daily rainfall and groundwater levels for 4 stations in the seasonally humid Upper Nile Basin of Uganda. The results of a linear regression of observed groundwater recharge against both total rainfall depth (representing daily rainfall) and heavy rainfall exceeding a threshold depth of 10 mm day-1 showed that for three stations, the magnitude of observed groundwater recharge was more strongly correlated to the latter (Figure 1). This indicates that groundwater recharge is biased towards heavy rainfall events rather than daily rainfall levels. These higher co-efficient of determination (R2) values were also associated with lower root mean square error values (rmse), increasing the reliability of these conclusions.



Figure 1. Scatter plots displaying the relationship (linear regression) between groundwater level rise during observed recharge events and total rainfall depth (representing daily rainfall) (dashed lines) and heavy rainfall exceeding a threshold of 10 mm day-1 (solid lines). (Source)

Furthermore, the regression of observed recharge and sum of total rainfall (dashed lines on Figure 1) to positive intercepts (158-189 mm) on the x (rainfall) axis (i.e. where the dashed lines cross the x-axis) supports previous assertions (12) that annual rainfall exceeding 200 mm is required for recharge to occur in tropical Africa (Owor et al., 2009).

More recently, the bias of groundwater recharge to heavy precipitation events has been supported by Taylor et al. (2013). Their 55-year record of groundwater levels in the Makutapora Wellfield aquifer of semi-arid central Tanzania highlights the highly episodic occurrence of recharge events under heavy rainfall, which interrupts the multiannual recessions in groundwater levels. Figure 2b displays the records cumulative recharge distribution, showing that a small number of high recharge events account for a disproportionately high amount groundwater recharge; indeed the top 7 seasons of rainfall account for 60% of the total observed groundwater recharge (Taylor et al., 2013). Furthermore, the record suggests that unless seasonal rainfall exceeds 670 mm (i.e. exceeds the third quartile), little or no recharge occurs (Taylor et al., 2013). For nearly 2/3 of the record no recharge is observed, yet after this threshold the groundwater recharge increases exponentially with increasing seasonal rainfall, furthering the point that recharge is restricted to infrequent events of high intensity precipitation. As such, overall Figure 2 shows that the observed relationship between seasonal rainfall and groundwater recharge is non-linear, as recharge is disproportionately restricted to this anomalously intense seasonal rainfall.



Figure 2. a) Observed groundwater recharge from groundwater-level fluctuations and rainy season (November-April) rainfall (shaded region shows seasons with a month of statistically extreme rainfall, i.e. >95th percentile), solid line=median; dashed line=third quartile; b) Cumulative contribution of annual recharge (ranked from highest to lowest) to the total recharge received at the Makutapora Wellfield 1955-2010. (Source)

A key restriction of both Taylor et al.’s (2013) and Owor et al., (2009) papers are the number of groundwater aquifers analysed. With only 5 separate groundwater aquifers analysed between the two papers, this left me with the question – is groundwater recharge biased to extreme heavy rainfall events not just in these aquifers, but across Africa, and indeed the rest of the tropics?

For example, with Owor et al.,’s (2009) study, the differences in slope (evident from differences in the axis scales for each wellfield) primarily reflect variations in the storage properties of the monitored aquifers, since the recharge surface environments (e.g. soil type) of the aquifers are all similar (Figure 1). This demonstrates that factors such as the geology of the groundwater aquifer may influence its ability to transmit heavy rainfall through into storage, and thus not all groundwater aquifers may be biased towards heavy rainfall events.

However, a recent paper hot off the press addresses this question exactly. In order to demonstrate the ubiquity in the bias of tropical groundwater recharge to intensive threshold-exceeding precipitation, Jasechko and Taylor (2015) related long-term records of stable isotope ratios of O (18O/16O) and H (2H/1H) in tropical precipitation at 15 pan-tropical sites (including Africa, Asia and the Americas) to those of local groundwater. In the tropics these ratios are strongly determined by site-scale precipitation intensities; high-intensity rainfall is comparatively depleted in heavy isotopes (18O, 2H) (Jasechko and Taylor, 2015). As such, the comparison of rainwater isotope composition for all rainfall events with the isotope composition of groundwater recharge-generating rainfall enables tracing of the rainfall intensities that produce groundwater recharge.

The results of the study were arresting: the comparison revealed that groundwater recharge in the tropics is near-uniformly (14/15 sites) biased to intensive monthly rainfall, commonly exceeding the ~70th precipitation intensity decile. Figure 3 shows that 14 out of 15 sites have mean groundwater δ18O values lower than the (i.e. groundwater depleted in 18O- and 2H relative to) long-term amount-weighted precipitation δ18O. This isotopic data confirms that groundwater recharge is biased towards 18O- and 2H- depleted rainfall – in other words, more intense rainfall.


Figure 3. Groundwater and long-term amount-weighted precipitation isotope compositions at 15 pan-tropical sites. 14/15 sites have 18O- and 2H-depleted groundwater relative to long-term amount weighted precipitation. Errors marks average +/- one standard deviations. (Source)

This analysis revealed that this bias is pan-tropical, occurring in a wide range of hydrological environments, for a variety of aquifer types, soil types. This confirms that a wide range of geological, climatological and land-use conditions are able to transmit intensive rainfall exceeding the median to shallow aquifers – thus the bias of recharge to heavy rainfall is not restricted to the few aquifers addressed in Owor et al.’s (2009) and Taylor et al.’s (2013) analyses.

For each site, Jasechko and Taylor (2015) matched the mean groundwater isotope compositions with those of the amounts-weighted precipitation under varying precipitation intensity thresholds, in order to trace the threshold precipitation intensities required to produce groundwater recharge. The intersection of groundwater δ18O and the amount-weighted precipitation δ18O under varying precipitation intensities provides an estimation of precipitation intensity thresholds required to initiate groundwater recharge. At all the sampled locations across tropical Africa, the groundwater isotope compositions were consistent with monthly rainfalls exceeding the ~70th percentile intensity (i.e. the range of >30th to >90th percentile) (within 1 standard deviation) (Figure 4). The authors conclude that under the primarily humid conditions experienced at the sites, these potential intensity thresholds correspond to rainfall intensities of ~100-300 mm/month, suggesting that rainfall intensities below these thresholds do not contribute substantially to groundwater recharge. *



Figure 4. Long-term amount-weighted precipitation δ18O  using data exceeding precipitation intensity thresholds and local groundwater δ18O values. (Source)

Overall, under the predicted regime of decreasing frequency of low- and medium-intensity rainfall events, and increasing frequency of very heavy precipitation events, groundwater may be a plausible solution to the negative effects this will have on the reliability of river discharge and soil moisture. Groundwater resources are better distributed than surface waters and account for over 90% of accessible freshwater worldwide (Shiklomanov and Rodda, 2003). They present low-cost strategies to adapt to changing freshwater availability and demand (Owor et al., 2009), through strategies such as groundwater-fed irrigation and sustenance of domestic and industrial water suppled (Jasechko and Taylor, 2015).

However, there are a number of important uncertainties with the three papers reviewed in this post. A key uncertainty overall concerns whether soil infiltration capacities are able, in practice, to transmit the modelled increases in recharge generated by heavy rainfall. Although Jasechko and Taylor (2015) demonstrated the ubiquity of soil types that are biased towards heavy rainfall in groundwater recharge, the relationship between precipitation and groundwater recharge remains poorly resolved in many regions due to a lack of long-term observational data (Taylor et al., 2013).

Furthermore, substantial uncertainty remains as to whether potential rises in groundwater recharge under climate change will be offset by increased evapotranspiration associated with warmer atmospheres (Owor et al., 2009; Kingston et al., 2009). The observed thresholds of the three discussed papers reflect the requirement of intense rainfall to overcome high rates of PET that prevail in the tropics in order to generate recharge (Taylor et al., 2013). The conversion of precipitation into groundwater recharge at low latitudes is constrained by continuously high PET rates (Jasechko and Taylor, 2015), estimated locally to be 160 mm month-1 during the monsoon season in the Makutapora Wellfield in Tanzania for example (Taylor et al., 2013).

A last point is that the resilience of groundwater to climate change can be undermined by other influences on groundwater storage such as human overuse, changes in the total volume of precipitation, and land-use change. Taylor et al. (2013) observed that rates of groundwater level decline in Tanzania have increased substantially from ~0.5 m yr-1 (1955-1979) to ~1.7 m yr-1 since 1990. As Figure 5 shows, multiannual declines in groundwater can be strongly linked to increases in monthly groundwater abstraction, that has increased from 0.1 to 0.9 million mover the 55 year period in order to supply potable water to the national capital, Dodoma (Taylor et al., 2013).



Figure 5. 55-year record of groundwater levels (top panel) and monthly groundwater abstraction (lower panel) from the Makutapora Wellfield, central Tanzania. (Source)

*An important uncertainty associated with this conclusions is that the analysis assumed all intensive rainfalls under increasingly higher decile thresholds contribute equally to groundwater recharge – however, Jasechko and Taylor (2015) point out that some (limited) data reveal that the proportion of very heavy, statically extreme rainfalls (e.g. >95th percentile) converted to recharge can be substantially greater than less intensive rainfalls. As such, the problem here essentially lies with the decile system, which groups together rainfall intensities above each decile.

Thursday, 3 December 2015

Precipitation in Africa – A Warming World

I trust you are well and ready to get stuck into some important issues! In this blog post I’m going to be discussing one of the major impacts of predicted climate change on African water resources – that on the frequency and intensity of rainfall. As we saw last week, there is a strong coupling between rainfall and river discharge in Africa, and as such, changes in the rainfall regime of Africa are going to have serious impacts on river discharge and thus water availability.

The Clausius-Clapeyron Relation

There is one scientific relationship that is particularly important when thinking about the impact of climate change on precipitation in Africa – the ‘Clausius-Clapeyron Relation’ (Figure 1). This dictates the increasing ability of air to retain water vapour with increasing temperature; as such, warm air has a much greater capacity to hold moisture than cold air. A very well determined value is the change in water-holding capacity of the atmosphere, governed by the Clausius-Clapeyron equation, of about 6.5% K-1 (Trenberth et al., 2003; Owor et al.,2009).


Figure 1. The Clausius-Clapeyron Relation.

The mean annual temperature rise over Africa is likely to exceed 2 ° C relative to the last 20th century by the end of this one – and that’s just under the medium scenarios (Niang et al., 2014). With this temperature rise, the air is going to be able to hold more moisture before it reaches dew point, condensation forms and precipitation occurs, as dictated by the Clausius-Clapeyron Relation. This leads to heavier, more extreme rainfall – so we move to a situation with a greater number of heavy, extreme precipitation events (i.e. those in the uppermost quartiles of the rainfall distribution) (Trenberth et al., 2003; Owor etal., 2009). Allan et al (2010) identify a 60% increase in the frequency of the wettest 0.2% of rainfall events per K warming. Furthermore, because the heaviest rainfall events usually deplete air of all its moisture, this rise in extremely heavy rainfall events is necessarily joined by a fall in the number of low and medium intensity rainfall events, or an overall decrease in the frequency of rainfall events (Trenberth et al., 2003; Owor et al., 2009). This change in the distribution of rainfall will result in more variable river flows and soil moisture (Owor et al., 2009).

There is already evidence that extreme precipitation changes over eastern Africa, such as droughts and heavy rainfall, were more frequent during the last 30-60 years (e.g. Williams andFunk, 2011; Lyon and DeWitt, 2012). Projected increases in heavy precipitation over the region have been reported with high certainty in the SREX (Seneviratneet al., 2012), and Vizy and Cook (2012) indicate an increase in the number of extreme wet days by the mid-21st century.

Basically, this means there are fewer rain events, and the ones that do occur are more extreme. But the bad news doesn’t stop here I’m afraid. This relationship is exponential, which means that the higher the starting temperature, the greater the increase in the air’s ability to hold moisture with that 2° C rise in air temperature – NOT good news for the tropics! The colder starting temperatures of the temperate regions – like us here in the UK! – leads to a comparatively smaller increase in the ability of air to hold moisture with global warming. So, despite greater increases in surface temperatures projected for higher latitudes, the exponential relationship with temperature in the Clausius-Clapeyron relation means there is a bigger absolute increase in moisture amount able to be held in the air in lower latitudes compared to higher latitudes (Trenberth et al., 2003; Owor et al., 2009).

Current Evidence

Allan and Soden (2008) use multiple modelling simulation and satellite observations to examine the response of tropical precipitation events to naturally driven changes in surface temperature and atmospheric moisture content. As such, the studies used natural climate variability induced through El Nino and La Nina events to successfully demonstrate a direct link between higher frequencies of very heavy precipitation with warm El Nino events (a statically significant relationship, according to Allan et al., 2010), and lower frequencies with cold La Nina events. Ultimately this supports the argument that temperature rise in Africa will lead to the increased frequency of extremely heavy rainfall at the expense of light rainfall. The studies also found satellite observed amplification of rainfall extremes under this climate change regime to be several times larger than that expected from the Clausius-Clapeyron Relation and at the upper limit of most responses in model simulations, implying that model projections of future changes in rainfall extremes in response to anthropogenic global warming may be currently underestimated (Allan and Soden, 2008; Allan et al., 2010).

Figure 2 describes these trends in the frequency of light- to heavy- precipitation events with time for GPCP data (Allan and Soden, 2008). Taking the straight horizontal line as the mean, it can be seen that there is an increase in frequency above the mean for rainfall events in >60 precipitation percentile (i.e. the heaviest rainfall events), and a decrease in frequency for rainfall events in the smaller precipitation percentiles.

Figure 1. Trends in the frequency of precipitation events with time for GPCP data. Source: Allan et al., 2010.

Few! So there we have - with temperatures rising over Africa, there's going to be an increase in extreme, heavy rainfall events, and a decrease in low- to medium- rainfall events. There is already satellite and model evidence for this. Not only this, but the effects of this are going to be greater in the tropics than in higher latitudes.

Thursday, 26 November 2015

Variation in the Climate Change Signal of the Discharge for the River Mitano Part 3: Hydrological Model, Downscaling and Lack of Data as Sources of Uncertainty

Uncertainty in hydrological model structure and calibration, and parameterisation uncertainty

Uncertainty in the hydrological model can result during its calibration. For example, Kingston and Taylor (2010) experienced occasionally very poor month-to-month performance of the model, with substantial monthly flow events either missing or erroneously introduced in the modelled time series. They attributed this to discrepancies between the observed (station-based) precipitation data within the Mitano Basin and the CRU TS 3.0 data set used for calibration of hydrological model – in other words, the gridded rainfall data over the Mitano Basin that the model was run with was of dubious accuracy (Kingston and Taylor, 2010). This poor month-to-month performance remained following autocalibration routines and manual calibration of the most sensitive parameters, despite close matching of the flow duration curves (Kingston and Taylor, 2010). Even in the context of possible errors in the observed discharge data of +/- 15% (Mileham et al., 2008), the errors produced were large, showing that some uncertainty pertaining to model calibration will always remain.

Uncertainty in the parameters of the hydrological model (i.e. model parameterisation) is a possible cause for additional uncertainty in climate change projections of freshwater. However, Kingston and Taylor (2010) undertook an uncertainty analysis through manually varying the 10 most sensitive parameters, including evapotranspiration and soil water capacity, in their hydrological model for the Mitano Basin. This indicated that the effects of hydrological model parameter uncertainty on simulated runoff changes are small when compared to those associated with choice of GCM, climate scenario and PET algorithm, in agreement with much literature (Cisneros et al., 2014; e.g. Arnell, 2011Lawrence and Haddeland, 2011). However this was a limited and subjective assessment only performed on the HadCM3 GCM, so greater parameter uncertainty may be present for GCMs with more extreme climate scenarios.

Uncertainty in downscaling

The estimated impacts of climate change on freshwater resources can be very dependent on the approach used to downscale the climate model data, as demonstrated through systematic evaluations of different methods (Quintana Segui et al., 2010). In fact the range in projected change between downscaling approaches can be as large as the range between different climate models (Chen et al., 2011). Downscaling methods may also encompass assumptions, for example, Kingston and Taylor (2010) downscaled monthly climate data to a daily scale suing the ClimGen weather generator, which is based on the assumption that the pattern of climate change is relatively constant for a given GCM (Arnell and Osborn, 2006) – yet it is questionable if this is always the case.

Uncertainty due to lack of data

As mentioned previously with PET estimation, the projections of freshwater availability are limited by the quality of input data. In the case of Kingston and Taylor (2010), neither temperature nor evaporations data were available for the Mitano basin, so that the closest available observations were from the town of Mbarara, approximately 50 km to the East of the catchment. This would have introduced uncertainty into their models. Furthermore, the Mitano hydrological model was calibrated for the 1961-90 baseline period using daily river discharge data from only a single gauging station within the catchment – where data from more gauging stations would have been preferable (Kingston and Taylor, 2010). Lastly, in validating the hydrological model, the closest rainfall station to the Mitano River Basin in the CRU TS 3.0 data set became unavailable from 1996 onwards, with the new nearest available rainfall station located approximately 115 km from the basin – potentially being the cause of poorer performance of the validation period for the hydrological model compared to the calibration period (Kingston and Taylor, 2010).

Conclusions


This review includes the evaluation of uncertainty due to choice of GCM, estimation of PET, hydrological model structure and calibration, and limitations of the data used. The main outcome is the major dependence on the GCM used, with estimation of PET being another major source of uncertainty. Climate sensitivity is another source of uncertainty (Kingston and Taylor, 2010) which has not been explored in this post, although Cisneros et al. (2014) argue that the relative weightings given to different emissions scenarios are typically less important in determining the distribution of future impacts than the initial selection of GCMs considered. For the near future at least, uncertainty in modelling future precipitation and freshwater resources under climate change will always have some uncertainty, so climate change scenario runs should always be evaluated within the context of all this uncertainty.

Thursday, 19 November 2015

Variation in the Climate Change Signal of the Discharge for the River Mitano Part 2: Different Estimations of Potential Evapotranspiration (PET) as a Source of Uncertainty

The estimation of the PET climate change signal is an important contributor to uncertainty in prediction of freshwater resources, particularly in regions where precipitation (P) is closely in balance with PET (Kingston et al., 2009), and there is a seasonal switch between P-PET deficit and surplus (Kingstonand Taylor, 2010).

Over 50 different methods of estimating PET exist (Lu etal., 2005), which each encompass different rates of change in PET for water resources. The Food and Agricultural Organisation of the United Nations (FAO) recommends use of the Penman-Monteith equation (Allen et al., 1998), however, in many locations (with certain climatological environments and where empirical calculations are robust) insufficient data prevent its use and compel application of simplified, empirical methods requiring fewer input data (Kingston et al., 2009). Priestley-Taylor is a widely used simplification of Penman-Monteith, and Hargreaves is the method recommended by the FAO in absence of sufficient data to calculate Penman-Monteith (Kingston et al., 2009). Therefore the differences between these three methods is particularly important for examining differences in projections of freshwater in practice.

Kingston et al.’s (2009) study compares the global response of 6 commonly used PET methods within hydrological models to a 2 °C rise in global mean temperature (relative to a 1961-90 baseline) under different GCMs. The results, displayed in Figure 1, show that different methods of estimating PET can produce markedly different climate change signals, and are thus an important source of uncertainty in freshwater predictions.



Figure 1. Annual PET for (a) the 1961-1990 baseline and (b-f) the assigned 2 °C climate change signal (scenario minus baseline), grouped by GCM. Ham=Hamon; Har=Hargreaves; PM=Penman Monteith; PT=Priestley-Taylor; BC=Blaney-Criddle; JH=Jensen-Haise. (Source)

Although PET increases at all latitudes for all PET methods and GCMs for the 2 °C climate scenario, differences of over 100% (> 200 mm in the tropics) exist in the climate signal between PET methods, for all GCMs (Figure 4) (Kingston et al., 2009). Maximum absolute differences occur at the same latitudes as the peak PET climate change signal (Kingston et al., 2009) – often coinciding with tropical African regions. Notably, although Priestley-Monteith, Priestley-Taylor and Hargreaves are generally closely grouped, they can still exhibit differences of over 60 mm at certain latitudes for certain GCMs (Kingston et al., 2009).

Table 1 shows that the mean range (variation) between PET methods and GCMs is higher for East Africa, where the Mitano Basin is located, than the Mediterranean and Southeast Asia. The key point here is that in each region, the choice of PET methods adds substantial uncertainty to the existing uncertainty associated with climate change signal between GCMs, and is of a comparable magnitude. Furthermore, in the case of East Africa and Southeast Asia, the choice of PET method for certain GCMs can actually determine the direction of projected changes in water surplus entirely (Kingston and Taylor, 2010).



Table 1. Statistics regarding percentage change in water surplus for three selected regions. (Source)

At the basin-scale, Kingston and Taylor (2010) compared baseline-scenario PET ratios for the three commonly used methods of calculating PET. The results are displayed in Table 2, and show that substantial differences exist in the PET climate change signal between the three methods for the HadCM3 2 °C scenario, with this difference most apparent during May, where the increase in PET ranges from 9-17% between methods (Kingston and Taylor, 2010).



Table 2. Percentage Change in monthly PET from baseline to the assigned 2 °C HadCM3 climate scenario. (Source)

This method dependence of PET estimation can be attributed to two factors. The first is the inclusion of different meteorological variables in each method (Kingston et al., 2009). For example, whilst Penman-Monteith is based on net radiation, temperature, wind-speed and vapour pressure), Priestley-Taylor is based on solely on net radiation and temperature (Priestley and Taylor, 1972). Figure 1 shows that the two methods’ signals diverge strongly at latitudes encompassing large arid or semi-arid zones (e.g. 10-30 °N), which suggests the absence of humidity data in Priestley-Taylor is critical in their differences (Kingston et al., 2009). The second factor is the different empirical formulation of each method, demonstrated through comparison of the Hamon and Blaney-Criddle signals in Figure 1; both methods are based on mean monthly temperature and day-length, yet have markedly different climate change signals (Kingston et al., 2009). The reliability of PET methods is further dependent on the reliability of input climate data – for example, compared to the relatively high confidence that can be placed in data such as observed temperature, Penman-Monteith is subject to data collection uncertainties such as the reliability of specification of canopy conductance and cloud cover (Kingston et al., 2009).


Ultimately, differences in the representation of evaporation (and snowmelt) processes lead to hydrological model structural uncertainty on projected changes, which can be substantial (e.g. Hagemannet al., 2013; Schewe et al., 2013). In some regions (e.g., high latitudes; Hagemann et al., 2013) with reductions in precipitation (Schewe et al., 2013), hydrological model uncertainty can be greater than climate model uncertainty—although this is based on small numbers of climate models (Cisneros et al., 2014). Incorporating uncertainty in hydrological model structure would have increased the range in projected impacts (Figure 3) on the Mitano Basin catchment scale even further.

Thursday, 12 November 2015

Variation in the Climate Change Signal of the Discharge for the River Mitano Part 1: Choice of General Circulation Model (GCM) as a Source of Uncertainty

For the next 3 blog posts I shall be discussing uncertainty in predicting freshwater resources under climate change, and the various sources of uncertainty, with particular reference to the River Mitano in Uganda. There is a projected intensification of the global hydrological cycle over the 21st century under climate change (strong likelihood; Bates et al., 2008),  yet substantial uncertainty remains regarding the impacts of this intensification on water resources. To predict rainfall, we take evidence from a multiple GCMs, downscale them using an RCM, apply that evidence to a hydrological model which then produces multiple estimates (uncertainty) in future hydrological flows and stores. Substantial uncertainty is introduced at each stage of this process, and ultimately results in huge variations in river discharge projections. In this blog post I shall be illustrating this variation in projections of precipitation and river discharge for the Mitano River Basin in Uganda, and discussing the associated sources of uncertainty. However it is important to note that uncertainty in climate change projections of precipitation does not just affect Africa – it is a global issue.

Kingston and Taylor (2010) modelled the response of the River Mitano in Uganda to a prescribed warming of 2 °C in global mean temperature using seven GCMs. Under these 7 GCMs, the rise in annual air temperature for the Mitano catchment ranged from approximately 1.8 °C for the NCAR GCM to 2.7 °C for the MPI GCM. However, Figure 1 shows that the differences between GCMs were much greater for precipitation, with projected changes in mean annual precipitation over the basin ranging from a 23% increase (NCAR) to a 23% decrease (CSIRO). On a monthly basis, some GCMs project greater (lower) precipitation in all months whereas other project a mixture of increasing (decreasing) precipitation over the course of the year (Figure 1b). As such, uncertainty in temperature projections occur mostly from divergent warming scenarios, rather than GCMs (Cisneros et al., 2014). There is greater confidence in the prediction of temperature than there is in any prediction of rainfall in the future.



Figure 1. Graphs illustrating the differences between scenario and baseline a) temperature and b) precipitation, under the assigned 2 °C warming scenario for 7 GCMs. (Source)

The simulated discharge for the Mitano under the 2 °C scenario shows these substantial disparities between GCMs, with little consistency in either the magnitude or direction of change on either a seasonal or annual basis (Figure 2) (Kingston and Taylor, 2010). For example, the CSIRO GCM projections result in reductions in river discharge at 2 °C that are greater than those from the HadCM3 GCM at 6 °C; whereas the NCAR GCM projections result in increased river discharge in all months that is double that of the baseline in some situations (Figure 2a) (Kingston and Taylor, 2010). In fact, the seven GCMs used here are drawn from an even larger population of 23 CMIP-3 GCMS, which, if included, would likely increase the variation in predicted discharge even further (Kingston and Taylor, 2010).


Figure 2. Climate change signal in a) monthly; b) annual discharge of the River Mitano under the assigned 2 °C warming scenario for 7 GCMs. a) (Source) (IPCC); b) (Source)

Resolution of the climate change signal for each GCM derived from temperature and precipitation separately reveals a comparative consistency in the temperature signal between GCMs (Figure 3a). In contrast, the similarity of the precipitation signal (Figure 3b) to the River Mitano’s monthly discharge signal (Figure 3c) clearly indicates that changes in precipitation are the dominant driver of changing river flow for the 2 °C scenario. Furthermore, the variation in the precipitation climate change signal between GCMs is far greater than that for the temperature signal, with different GCMs giving rise to both positive and negative changes in discharge in all months (Kingston and Taylor, 2010). Thus variability between GCMs in the precipitation climate change signal, rather than temperature, is the main source of variability between GCMs in the discharge signal for the Mitano (Kingston and Taylor, 2010). This is despite East Africa being identified as a region where there is relatively good agreement between precipitation projections (Christiansen et al., 2007). This is in agreement with several other authors, including Cisneros et al. (2014) who argue that seasonal distribution of change in streamflow varies primarily with the seasonal distribution of change in precipitation, which in turn varies between scenarios. Both the patterns of change and uncertainty are driven largely by projected changes in precipitation.


Figure 3. a) Monthly river discharge climate change signal for the Mitano; b) temperature-only climate change signal for Mitano; c) precipitation-only climate change signal. All for the River Mitano under the assigned 2 °C warming scenario for 7 GCMs. (Source)


The results of Kingston and Taylor’s (2010) modelling of future river discharge for the Mitano Basin clearly shows that the GCM used to model the effects of climate change on river discharge is a highly important source of variation (i.e. uncertainty) between projections of precipitation and thus freshwater resources. As such, single-GCM evaluations of climate change impacts on river discharge are likely to be wholly inadequate and misleading (Kington and Taylor, 2010). However, there are other important sources of uncertainty in predicting future freshwater resources, that I shall discuss further in the next posts.