Report: Groundwater depletion clouds Yemen’s solar energy revolution

Summary

Warnings have long been sounded about Yemen’s water security. The country has a per capita water availability roughly 1.3% of the world average, and a groundwater extraction rate well in excess of its recharge rate.¹ The issue of long-term water availability, however, has been eclipsed by the immediate humanitarian needs in the country, now torn apart by years of war. More than 10 million people – a third of the population – are at risk from famine, with 47,000 expected to experience it in the first half of 2021.

Amidst the horror of war, however, a positive story has begun to emerge. The country, which faced months of blackouts as the national grid collapsed in the early stages of the conflict, and is subject to crippling fluctuations in diesel markets for its power, has started to embrace solar energy. Markets for solar panels are booming as anyone with the capital invests in solar in order to meet the basic needs of their households. Solar has now spread to health, education and agriculture.

Agriculture in Yemen is dependent on diesel for the extraction of groundwater and successive crises in fuel markets – as they shape and are shaped by the conflict – have taken a heavy toll. Solar power has the potential to break this devastating cycle and increasingly, this is becoming a reality. Amid reports of unproductive land becoming fertile again, international development agencies have invested in solar power, the government is tendering for vast projects, and the private sector is heralding its deployment of this life saving technology. The take-up of solar by individual farmers also appears to be significant.

But the ‘solar revolution’ in Yemen is not without its dangers. CEOBS has used an approach based on satellite remote sensing to assess changes in groundwater levels, information about which has not been analysed since the conflict disrupted the country’s water well monitoring capacity. In this report, Leonie Nimmo and Eoghan Darbyshire analyse groundwater change in Yemen, incorporating insights from expert interviews alongside data on energy markets, agriculture, the armed conflict and rainfall. We conclude that a significant drop in groundwater since 2018 is likely a result of the spread of solar in agriculture, and argue that interventions are required on multiple levels and by all stakeholders to halt severe groundwater depletion.

The research

CEOBS has sought to assess changes in groundwater levels in the west of Yemen – home to 90% of the country’s population² – using satellite measurements of the Earth’s gravitational field and soil moisture.³ We adopt a holistic analysis by combining this with expert interviews and complimentary datasets: precipitation, diesel price, agricultural production, conflict intensity, solar power data, satellite imagery, and vegetation data.

Taking each dataset in turn, the picture in western Yemen as a whole is examined, beginning with background information on the relationship between groundwater, energy use and war, and the historical agricultural context. In part two we assess the heterogeneity of the western Yemen groundwater trend by clustering regions with similar time-series. This is followed by detailed analysis of groundwater in the central and northern highland plains, and the coastal Tihamah region.⁴ In part three we review the prospects for future groundwater monitoring and take stock of the solar stakeholders in Yemen. Recommendations for actions on all levels to avert precipitous groundwater decline are presented in part four. These are summarised in Table 1 below.

Key findings

• Groundwater in western Yemen is at its lowest level since satellite records began in 2002,⁵ in spite of some recovery in the early years of the conflict and above average levels of rainfall in recent years.

• We hypothesise that these drops are driven by the spread of solar power, which is decoupling the historical relationship between diesel markets and groundwater pumping. In the early years of the conflict this relationship constrained the extraction of groundwater,⁶ resulting in agricultural losses that negatively impacted food security.

• In some parts of the country a continuation of these trends may result in the accessible groundwater running out, or becoming inaccessible for those most in need. Furthermore, associated problems of land subsidence and sea water intrusion will increase.

• There is a need for further expert technical analysis of the groundwater situation in Yemen, combining remote sensing with existing hydrogeological data, and renewed well monitoring with local stewardship, where feasible.

• Averting severe groundwater decline will require action to be taken on multiple levels by all stakeholders, from civil society organisations to international agencies. Addressing water rights and access, and supporting the development of sustainable livelihoods, are key.

• There is a need to improve the exchange of knowledge in Yemen about groundwater resources in general, and the impact of solar-powered water pumping in particular.

• Civil society in Yemen can play a key role in groundwater stewardship, and should be supported with data and appropriate technologies.

• An important aspect of future sustainable groundwater management may lie in reviving the country’s ancient water harvesting systems, and utilising extant traditional management systems and structures.

• The massive potential for solar power in Yemen should be managed with awareness of the potential impacts on groundwater and the problems associated with electronic waste.

• Data on the scale of solar-pumping across Yemen is limited, but it is urgently needed to understand and manage the associated risks.

Recommendations

Table 1. Overview of recommendations.

Contents

Part 1. The national picture

1.1 Groundwater levels in western Yemen

1.2 Background: groundwater, energy use and war

1.3 Historical context: agricultural change

1.4. Factors affecting groundwater abstraction

1.4.1 Diesel and conflict

1.4.2 Solar power upsurge

1.4.3 Armed violence

1.4.4 Agriculture

Part 2. The regional picture

2.1 Identifying regions

2.2 Regional groundwater trends

2.3 Central highland plains, including Sana’a

2.4 Southern Tihamah

2.5 Northern highland plains

Part 3. Taking stock

3.1 Implications of continued falling groundwater

3.2 Climate change

3.3 Characterising the groundwater of Yemen

3.4 Electronic waste

3.5 Solar stakeholders

Part 4. Interventions

4.1 Groundwater monitoring

4.2 Knowledge exchange

4.3 Sustainable livelihoods

4.4 Environmental peacebuilding

4.5 Private sector

Conclusion

Appendix

Part 1. The national picture

1.1 Groundwater levels in western Yemen

Figure 2 shows groundwater and precipitation changes in western Yemen from 2002 to 2020. As rain recharges aquifers there is a relationship between these variables, though other factors are also at play. We have also plotted the recorded changes in groundwater levels against our crude estimate of the changes that would have been expected as a result of precipitation.⁷

Figure 2. Groundwater and precipitation changes in western Yemen. Click left/right arrows to toggle between precipitation and expected groundwater plots.

During the first half of the conflict and immediately prior to it (2013-2017), groundwater levels were recovering in western Yemen. From 2018 onwards, they have been in decline, reaching their lowest point on the satellite record by 2019 and falling since. This is in spite of rainfall levels that were well above normal for the last 20 years.

We hypothesise that the decline in groundwater since 2018 uncovered through this research can be attributed to a decoupling of the dependency on diesel for extraction and a new pattern of abstraction that is going largely unchecked, made possible through the spread of solar power.

In addition to over abstraction, pollution remains a sister threat to groundwater in Yemen. Pollution sources include the oil sector, improper waste treatment,  industrial and medical sources, and agricultural run-off of banned pesticides. Furthermore, as groundwater levels fall the increased intrusion of seawater into aquifers turns them salty. We do not explore groundwater pollution in this report, but it is a topic deserving of further attention.

Return to Contents

1.2 Background: groundwater, energy use and war

Prior to the onset of war in Yemen it was clear that the rapid and unregulated proliferation of boreholes was causing water tables to drop dangerously in some areas of the country. With an estimated 100,000 mostly illegal wells in the capital Sana’a and surrounding area alone,⁸ the city was slated to be one of the first capital cities in the world to run dry. Though such an apocalyptic scenario has so far not played out, water wells being drilled there today are up to an extraordinary 1.2 km deep, though on average they are 350 m.⁹ To put that in context, the average depth of a rock well in the United States is 76 m. In the Sana’a basin, water that accumulated eight thousand years ago has been exhausted in 30 years, according to Yemeni water engineer Taha Al-Washali.¹⁰

Yemen’s water well monitoring system has fallen victim to the war,¹¹ leaving a near-complete void of information about the current state of the country’s groundwater resources.

At the outset of the conflict, agriculture accounted for 90% of groundwater use, so understanding the agricultural context is key to determining the drivers of – and by extension, remedies for – groundwater depletion in the country. However, whilst 73% of the population were dependent on agriculture for survival in terms of household incomes, the country only produced 10% of its staple foods. These are skewed statistics for a highly arid and poor country, and signalled a vulnerability to famine that was apparent before armed conflict engulfed it.

Yemen has been dependent on diesel-powered groundwater extraction since the country restructured its agricultural system from the 1970s onwards with the guidance of the international community. The war had an immediate and extreme impact on energy supply: the national grid collapsed and the price and availability of diesel fluctuated massively.

In response to the dire state of hydrocarbon-based power supply, the population has wholesale turned to solar power for household needs, and it appears to be increasingly significant for agricultural production. But the question of how solar-powered groundwater extraction is affecting the country’s groundwater resources is not being properly addressed. Without the limiting requirement of fuel inputs, solar has the potential to significantly exceed diesel-powered extraction.

The risks of over-abstraction of groundwater when solar-powered technologies are deployed were recognised by the Food and Agriculture Organization (FAO) of the United Nations in 2018:¹²

“SPIS [Solar Powered Irrigation Systems] – if not adequately managed and regulated – bear the risk of supporting unsustainable water use. Once the systems are installed, there is no cost per unit of power and thus no financial incentive for farmers to save on fuel or electricity for water pumping. This can lead to wasteful water use, over-abstraction of groundwater, and low field application efficiency.”

Return to Contents

1.3 Historical context: agricultural change

Until the 1970s, agriculture in Yemen was intimately linked to the country’s geography and climate. According to Yemen agricultural specialist Anthony Milroy,¹³ farmers provided the links in a chain allowing rainwater to flow from ancient, intricate terraces in the highlands to the flood-based spate agricultural systems of the fertile coastal regions.^14,15 Milroy first went to Yemen as an employee of the British Overseas Development Administration in the 1970s. With an original mission to help Yemen develop its agricultural sector, he was taken by surprise by the highly sophisticated nature of Yemen’s traditional water management systems. “We didn’t have anything to teach them,” he says, “they had so much to teach us.” Spate irrigation is thought to have originated in Yemen 5,000 years ago.

It was not, however, a system that was recognised as valuable by the international development agencies of the 1970s. Government aid agencies as well as organisations such as the World Bank began to promote groundwater extraction as key for Yemen’s agricultural development. A neo-liberal model of development was pushed on Yemen as it was on the rest of the Global South, with an emphasis on foreign trade liberalisation and connecting subsistence producers – and consumers of food – to global commodity markets.¹⁶

Traditional grains were abandoned in favour of cash crops for export, whilst imported, cheap wheat from North America, Europe and Australia fundamentally changed Yemeni diets as well as its agricultural base. Meanwhile, in Milroy’s words, exotic crops such as bananas and mangoes effectively became the conduit via which Yemen’s groundwater was exported, with Saudi Arabia a key export market.

These shifts were enabled by groundwater abstraction, and diesel subsidies encouraged new diesel-powered wells. Spate irrigation systems were heavily modified by permanent water diversion structures.

The labour-intensive terraces of the highlands suffered from the 1970s onwards as the Yemeni labour force drained into oil-rich Saudi Arabia, a shift that also distorted the rural economy with remittance payments. Many terraces have now degraded and collapsed. In a matter of decades, the agricultural system switched from one that was entirely rain-fed to one that is critically dependent on groundwater for irrigation.

According to Milroy: “the impact of the abandonment of terraces and rainfed areas over the past five decades has been to massively exacerbate the surface water run-off, thus depleting the recharge in the mountains significantly and, as a direct consequence, causing continuing downstream flash flooding and loss of fertile, food cropping areas and yields in the Tihamah. This is the direct impact of the abandonment of rainfed water harvesting and storage after seven millennia of maintaining those traditions.”¹⁷

Figure 3. Haraz mountains, Yemen. Degradation of agricultural terraces is visible in some areas. Photo credit: Rod Waddington, Flickr, licensed under Creative Commons.

Irrigation projects from the 1970s onwards sought to divert and control the flow of water to facilitate agricultural development, with donors such as the World Bank favouring large scale, upstream-engineering projects. This led to conflicts over water resources due to the opposing demands of up- and downstream farmers. The construction of dams, reservoirs and canals for the purposes of irrigation also reduced the recharge of groundwater aquifers, making it more difficult and costlier to extract groundwater as levels dropped.

Deeper wells were needed to maintain existing levels of extraction, but these required more fuel, and therefore cash. Groundwater extraction has therefore not only been unsustainable but inequitable, benefitting wealthier farmers at the expense of poorer ones: “Small farmers are not able to take part in the “race to the bottom of the aquifer” according to the Environment, Conflict and Cooperation platform. Water sources are 90% privately-owned,¹⁸ which has increased income disparities. A 2007 World Bank report on Yemen’s water sector reform programme found that “In agriculture, ownership of a water source is correlated with higher income, and development of groundwater resources in recent years has contributed to growing income disparities as the better off have been able to capture the lion’s share of the resource.”

According to the International Committee of the Red Cross (ICRC), disputes over water in Yemen exist between tribes, urban and rural communities, domestic and agricultural users and up and down-stream areas. In 2014 it was estimated that 2,500 people died annually as a result of water-related conflict. “Wherever in Yemen you see aquifers depleting, you have the worst conflicts,” Abdulrahman al-Eryani, a former minister for water and environment, told a New York Times reporter in 2013.

However, a DfID-commissioned report states: “The danger with over-emphasising the incidence and importance of local water conflicts… is that it undervalues the local regulation and good management which does take place.”

According to Abdu Mohammed Seid of the Norwegian Refugee Council (NRC), a culture of mutual aid exists in Yemen. People support each other to the extent they are able, “from providing a meal to destitute people up to the level of drilling water boreholes or rehabilitating water supply schemes for community use.”¹⁹

Return to Contents

1.4 Factors affecting groundwater abstraction

1.4.1 Diesel and conflict

Yemen’s population is heavily dependent on agriculture, the majority of which is dependent on diesel. Since 2011, fluctuations in diesel prices and availability have led to social and economic upheaval; in turn the conflict has led to diesel market volatility.

When the Arab Spring erupted in 2011, Yemen’s people were the poorest in the Arab world but their president was one of the world’s richest.²⁰ Their revolution was colour coded pink, as a signal that they were peaceful and to denote an organisational stage in the uprising. Protests were met with state violence. In December that year, a four day, 320 km, peaceful march from Taizz to Sana’a delivered 100,000 protestors to the capital.

By 2014 the outcomes of a National Dialogue Conference had failed to deliver federalist divisions of territory or a cabinet power rebalance that was acceptable to either the Houthis in the north or separatists in the south. The price of diesel doubled in July as subsidies were removed, with an immediate impact on the ability of the population to access water, and for agricultural communities to farm. The decision was expected to push an additional half a million people below the breadline.

In August, Houthi-organised protests against the subsidy withdrawal drew tens of thousands of supporters. In September 2014 the Houthis took control of Sana’a, followed, in January 2015, by their seizure of the presidential palace. In March 2015 Saudi Arabia formed a coalition of states in opposition to the Houthi power grab and began an aerial assault on the country that, as of March 2021, continues.²¹

Figure 4 shows the 2014 spike in diesel prices, followed by around a ten-fold surge in the first half of 2015. This is likely to have been in response to the launch of the coalition offensive. In July the Houthis, in financial crisis, cut the remaining subsidies on fuel, which is evident in Figure 4 as a second spike in 2015. This was a cruel irony given the part that subsidies played in the Houthis’ rise to power.

Irrespective of spikes, diesel prices have been on an upward trajectory since the removal of subsidies in 2014. Although Yemen produces oil, it has very limited refinery capacity, and so exports crude oil and imports refined oil products. The blockade of Yemeni ports has therefore increased fuel price volatility.²²

In Figure 4 we have plotted data on diesel prices against the level of man-made light at night in western Yemen, as measured by the VIIRS satellite instrument. There is a precipitous drop in the night lights coincident with the increase in diesel price and the collapse of the national energy grid in March 2015. The nighttime lights data is a proxy for energy use.

Figure 4. Western Yemen nighttime lights and diesel prices.

Figure 5. Change in the luminosity of nighttime lights between 2014 and 2019 over Yemen and southern Saudi Arabia.²³

Figure 5 shows the spatial distribution of the nightlights across the country and the changes between 2014 and 2019. The nightlight decline is clear in all major cities except Marib, where the population swelled with people displaced by war.²⁴ Also clear is the difference in development between Yemen and Saudi Arabia – evident before the conflict but exacerbated further since.

As well as its impact on the price of diesel, the conflict has impacted the availability of diesel across the country due to a blockade of imports and critical disruption to the transport networks. “All goods are transported by road, leaving many remote communities at the mercy of the quality and security of the roads connecting them to the rest of the country,” wrote Yemen commentator Peter Salisbury in 2011, noting that those in small settlements with poor transport links were the country’s poorest people.

Much of the road network has been destroyed or damaged over five years of war, including key infrastructure such as bridges. According to estimates based on data compiled by the Civilian Impact Monitoring Project (CIMP), a single shelling incident of a road could affect up to 14,000 households, and 13 airstrike incidents on bridges affected over 100,000 households.

Where roads are functioning, checkpoints manned by various parties to the conflict have facilitated the unofficial taxation of goods, including diesel. The vulnerability of the population to disruptions in transport networks – in terms of food, fuel, and by extension, groundwater – helps to explain the catastrophic impacts of the conflict on livelihoods across Yemen.

Return to Contents

1.4.2 Solar power upsurge

Population turn to solar as the national grid collapses

Pre-conflict public electricity coverage in Yemen stood at about 50% – the worst in the Middle East. The sabotage of key power lines in June 2014 resulted in a national blackout, and highlighted the vulnerability of the country’s electricity system. The national grid collapsed as the country’s largest power plant went offline in March 2015, the same month as the coalition intervention was launched.

Infrastructure has been a causality – if not a target – of the coalition offensive,²⁵ and by the end of 2016 only 10% had on-grid access. Lack of electricity has profound impacts, not only for households and industry, but hospitals and vaccine cold chains, schools, and the supply and treatment of water.

Diesel-powered generators were the first substitute as national power went down, but as diesel markets fluctuated wildly those that could afford the initial investment turned to solar power. In recent years solar has become much more economically viable: the World Bank has estimated that costs dropped 80% between 2010 and 2017.

Sales of solar panels increased by more than 2,000% in two months, according to a small survey conducted by the Yemeni Small and Micro Enterprise Promotion Service (SMEPS), reported in May 2015. A World Bank-commissioned study published in 2017 estimated that around 75% of households in urban areas used solar power. The international trade journal PV Magazine reported that around 300 MW of power were installed between 2015 and 2017, a figure also based on a paper published by a local NGO. Arvind Kumar of the UN Development Programme (UNDP) told CEOBS that $2 billion of solar equipment has been imported in recent years.²⁶

Figure 6. Solar panels in Hodeidah.

Lack of evidential data

In spite of indications of an increase in the use of solar power, reliable, country-wide statistics are hard to come by. Figure 7 shows data from the available sources.

Perhaps the most robust current data is from a study by the Energy Access and Development Program (EADP), which estimated capacity using data from local surveys, the World Bank, the International Energy Agency and the UN Statistics Division. Unfortunately, this data only includes the years to 2017 and not the period since, which is when the use of solar is understood to have skyrocketed.

Data into 2019 is provided by the International Renewable Energy Agency (IRENA), although we assume this is only based on undocumented desk assessments as the notes state: “Due to the ongoing conflict in Yemen, no official government data sources were available since 2017”. We are highly sceptical of this data: The source reports the same capacity for 2018 and 2019, which is not borne out by anecdotal reports, and it is lower by a factor of two compared to the EADP data.

Figure 7. The available data on imports of solar equipment in Yemen.

An alternative approach is to look at relevant import data from UN COMTRADE. However, the only commodity code available to analyse for solar panels also includes low-cost light emitting diodes. It is therefore difficult to know to what degree significant jumps in 2018 and 2019 are attributed to solar power. Data for the commodity code that only denotes photovoltaic cells was unavailable.

The spread of solar

Hard data aside, it seems clear that from the small electrical shops experiencing booming demand in urban centres at the onset of the conflict, solar has spread throughout Yemen. Major development agencies have been investing in solar power infrastructure since at least 2018.²⁷ The government initiated a subsidy programme for solar power in agriculture in 2014, but this stalled with the conflict.

Stories of arid lands being bought to life accompany those of projects that have empowered women and transformed lives with the supply of affordable and environmentally-friendly energy. Finally, Yemen had a good news story. It also has something of value, with the highest average theoretical potential for solar power in the world. Figure 8 shows the spatial distribution of potential photovoltaic electricity production in Yemen. This is greatest in the highland areas, although even in the coastal areas the potential is significant – greater than the average theoretical potential in a selection of similar-sized countries – see Figure 8.

Figure 8. The spatial distribution of theoretical potential solar resource in Yemen. Data presented is the long term (1999-2018) yearly average of global horizontal irradiation (GHI) (kWh/m²), an output from the Solargis global solar model by the Global Solar Atlas. Comparative figures and global ranking are presented for countries with a similar population. GHI represents the theoretical potential, other practicalities such as land availability or use are not taken into account. Yemen also ranks highly (14th) for having the smallest proportion of its land required to produce its yearly electricity consumption – the so-called PV equivalent area.

Solar power is good news for Yemeni farmers: once the initial cost of the system is paid for, there are no ongoing fuel costs nor labour required for refuelling. In fact, “once connected to the power supply, the pumps are able to deliver continuous groundwater,” according to solar giant Suntech, which supplies local distributors in the country. This contradicts reports from the NRC’s Abdu Mohammed Seid, who said that the pumps were only run for seven hours.^28,29

The spread of solar power in agriculture is influenced by the local geography and topography, which defines the depths required of the wells or boreholes. In some highland areas the costs of solar power groundwater abstraction are prohibitive, says Seid.

What is remarkable about the public discourse on the ‘solar revolution’ in Yemen is that there is little mention of the dangers of continuous groundwater extraction, particularly in light of the grave concerns that have previously been raised about unsustainable groundwater abstraction in the country.

The UN International Groundwater Resources Assessment Centre recognises that over-extraction is the main environmental challenge of solar pumping, with “serious consequences for communities and the environment”. According to its website:

“It is crucial to have clear regulations and effective monitoring of groundwater pumping and levels in place. Without a proper groundwater monitoring network it is not possible to control pumping and sustain groundwater resources. Based on observations, assessments and predictions, water allocation agreements need to be made among users at the local level.”

Figure 9. Solar-powered groundwater pumping in Yemen. Images taken from the World of Engineering and Trade Company videos: 1, 2, 3, 4.

Return to Contents

1.4.3 Armed violence

The war has impacted groundwater in Yemen through a number of different mechanisms. Here we consider the direct impacts of armed violence on infrastructure and population movements. Economic activity and livelihoods are considered under the section on agriculture, below.

Armed violence and sectors indispensable to human survival

The Armed Conflict Location & Event Data Project (ACLED) collates data on armed conflicts and crises internationally. We have analysed the ACLED Yemen data and organised it according to three sectors critical for human survival: energy, water and agriculture. Figure 10 shows the spatial distribution of armed conflict events, with colour coding to depict the year of the event. No differentiation is made between which party to the conflict is associated with the events. The ACLED dataset does not provide complete coverage of all relevant events, nor will our filtering capture all of these out of the total of more than 60,000 incidents coded since 2015. However, the results presented here capture the spatial dynamics and allow a basic sectoral and temporal analysis.

Figure 10. Location and year of conflict events associated with sectors critical for human survival. Based on keyword analysis of the ACLED Armed Conflict Location & Event Data Project database. Click the left/right arrows to toggle between the sectors.

Figure 11 shows the percentage of total events that are associated with these three sectors. Energy and water associated events were most frequent in the early stages of the conflict. Towards the end of 2020 there was a resurgence in such events. Notable in terms of water infrastructure damage are the destruction of desalination plants, for example, in Hodeidah and Al-Mocha.

Figure 11. Time-series of the percentage of conflict events associated with sectors critical to human survival. Based on keyword analysis of the ACLED Armed Conflict Location & Event Data Project database – methodology outlined in the Appendix.

Population changes

At the time of writing (March 2021) there are more than 4 million reported displaced in Yemen and living within the country’s borders. Yemen’s people are “penned in by the ocean and the desert” with Saudi Arabia to the north. Its would-be refugees therefore remain largely Internally Displaced People (IDPs). It is hard to gauge the impact of such population movements on society and the environment. According to the Internal Displacement Monitoring Centre, “More than half of the country’s displaced households live in rented accommodation, but 35 per cent live in vulnerable conditions in informal settlements, collective centres, public buildings, tents and even out in the open.” Those living in IDP camps in the desert are acutely vulnerable to insecure access to water.

The demands on water resources will have shifted as a result of the population movements, but information on water markets is scant. The population in Yemen obtain drinking water from one of four sources: groundwater, via wells; piped water distribution networks (urban areas only); trucked water and bottled water. The vast majority of water is supplied privately.³⁰

In this study we recognise that 90% of water use is for agriculture, and have not incorporated analysis on population dynamics owing to data uncertainties. As such, the impacts of redistributed demands for drinking and hygiene water are not considered in detail in our regional analyses.

Return to Contents

1.4.4 Agriculture

Crop production in decline and food insecurity

During a conflict many of the variables upon which agricultural production is dependent are thrown into flux. In Yemen diesel is a crucial factor, as is the destruction caused by war. Conflict has also had a three-fold impact on the labour force resulting from death, displacement and unemployment. Alongside these factors, the cost and difficulties of obtaining agricultural inputs, and getting goods to markets, influence the viability of farming and crop choice.

Yemen’s fragmentation due to the conflict means that there are currently two parallel agricultural ministries. The Houthi de facto government in Sana’a has maintained a Ministry of Agriculture and Irrigation (MAI). This is active, and has continued to produce agricultural statistics that were being reported before the war. We reference these throughout the remainder of this report, but note here that they must be treated with caution. The internationally recognised government, based in Aden, also has an agricultural ministry, though this does not produce any agricultural statistics. Similarly, there are two energy ministries, based in Sana’a and Aden, and serving different areas of the country.

Figure 12 shows the approximate area and yields of key crops in western Yemen from 2007 onwards, relative to 2007, according to the statistics produced by the MAI in Houthi-controlled Sana’a. From 2013 or before, all crops decreased in terms of the area on which they were grown, with the exception of qat. For all except vegetables and qat, there were also decreases in yields.

It is hard to overstate the likely impact of this downward trend in agricultural production on livelihoods in Yemen, and what is an increasingly catastrophic famine situation. Whilst the country is estimated to import 90% of its food, 73% of the population are dependent on agriculture for their survival.

As the economist Amartya Sen pointed out with his entitlement theory of famine,³¹ it is not the decline in food availability that causes famine, but a breakdown in the ability to access food, for example, through buying or growing it. This chimes with reports of there being lots of food available in Yemen during the crisis. According to a 2017 BBC report: “In the frontline city of Taizz, under siege from Houthi rebels for two-and-a-half years, a stroll around one of the local markets reveals abundance. Huge pomegranates and oranges, fresh garlic, bananas, courgettes and tangerines. There are supermarkets full of produce too – fresh meat, eggs, steaks and ribs. So how on earth can this country be starving?” This observation was also made by Peter Salisbury in January 2020: “The problem is not that there isn’t food, the problem is that people don’t have the money to buy it.”³²

Whilst the statistics shown here reveal an overall decrease in the supply of food, they also show a breakdown in livelihoods that has been potentially more significant as a precursor to the current acute and chronic food insecurity in the country. Further factors that have affected the ability of the population to buy food include port blockades, public sector salaries not having been paid and inflation of food prices. In January 2021 a UN Panel of Experts noted that there has been “double-digit inflation and a collapsing currency, which has a devastating impact on the population”.

The crop that presents an exception to the rule of decreasing areas and yields is qat, which maintained a steady constant in both from 2012 until 2019, when yields significantly increased. Yields for almost all other crops also increased in 2019.

Figure 12. Changes in yields and the farmed areas of major crops, 2007-2020.³³

Qat

Estimated to account for 40% of all water use in the country, the narcotic qat is the crop to which Yemen’s water shortages are often attributed. Carl Bruch of the Environmental Law Institute (ELI) visited Yemen in 2005, and it was apparent even then that qat was a problem: “There were multiple stories of wells going dry due to over-extraction for qat.”³⁴

The crucial ratio of diesel price to crop price is more favourable to qat than other crops. This, coupled with the strong demand for it in local markets – up to 90% of men and 50% of women are estimated to chew daily – has meant that it has proved a resilient crop in the face of successive diesel crises.

The extent to which qat production has increased or intensified is a knotty problem and not one that we have been able to answer conclusively during this research. There are a number of reports claiming qat production to have increased: The Economist’s damning 2018 article on qat in Yemen claimed that its cultivation was increasing by 12% per year. According to the Global Environment Facility: “Many farmers shifted from coffee to qat in response to the ongoing conflict.” According to Salisbury, there are anecdotal accounts of qat becoming an even more important crop since the escalation of Yemen’s conflict.³⁵

Earlier work by CEOBS did not find a significant areal expansion in qat growth in the locations that it studied, however, we did note that production may have intensified as the crop can be harvested up to five times a year. Changes in growing conditions can affect the amount of groundwater that is required to produce the same amount of crop. An increase in abstraction could reflect unfavourable conditions rather than an increase or intensification of qat production. The MAI (Sana’a) statistics indicate that the area used to grow qat has stayed fairly constant since 2012, but there was a big increase in yields in 2019.

Through qat, conflict dynamics intersect with groundwater dynamics: Revenues from qat are of some significance to the Houthis. According to an interviewee, “Qat is being used in different ways to fuel the war. It is consumed by the fighters, who get some money, plus food and qat.”

Biomass changes

Biomass is the mass of vegetation in a given area. In an area of cropland, it is a product of agricultural or land management practices and environmental conditions, including the climate, soil and hydrogeology. Changes in biomass can occur because of changes in environmental conditions or changes in growing practices. To identify where there may have been changes in the latter, it is necessary to remove the effects of the former, that is, remove the effect of changes in precipitation, humidity, solar radiation etc. We estimate this here using a data product called the gross biomass water productivity.

Figure 13 indicates that in the cropland areas of western Yemen biomass significantly increased in 2019 and 2020; in 2020 it reached a level more than double that previously recorded since 2009. It was also a period of above average rainfall (see Figure 2). The gross biomass water productivity also increased, i.e. biomass was above what would be expected from the climatic conditions during the period. This is consistent with the MAI statistics and an increase in the use of groundwater for crop irrigation.

Figure 13. Biomass changes in areas of cropland in Western Yemen, 2009–2021.

Return to Contents

Part 2. The regional picture

2.1 Identifying regions

The previous section presented groundwater changes in western Yemen as a whole; here we investigate how the situation varies spatially. To do this, a k-means clustering approach was used to identify areas with similar patterns of change in groundwater levels since 2011, when internal unrest in the country began to escalate.³⁶ In Figure 14 below these are made up of pixels, which each represent approximately 50 km² on the ground.³⁷ Ten clusters were identified that each had a distinct pattern of change, or time-series. Note that these areas are not defined by any political boundaries, rather, the pixels selected form an area or cluster that exhibits an internally similar pattern of groundwater change.

Because the area covered by some of the clusters is less than the native spatial resolution of the GRACE satellite data,³⁸ these more localised results can only be treated as indicative at this stage. Furthermore, due to land-sea masking,³⁹ some agriculturally productive and politically significant coastal areas cannot be represented – those in the northern Tihamah and western Tuban delta, near Aden.

Figure 14 shows the clusters studied, colour coded as they appear in the rest of this report. Key data sets for three regions are presented in this section; the Appendix includes the remaining areas and a full methodology.

Figure 14: Yemen map showing the clusters with self-similar groundwater changes identified. Areas have been named based primarily on geography; the clusters were identified due to the pattern of groundwater change in each pixel.

Return to Contents

2.2 Regional groundwater trends

Figure 15 shows the groundwater time-series in each area. Table 1 summarises regional information on groundwater trends, aquifers, crops and the precipitation-recharge relationship.

Figure 15: Groundwater time-series in each of the regions studied. The divergence in time-series after 2011 is to be expected, given this was the period used in the k-means approach to define the regions.⁴⁰

During the earlier years of the conflict (2014-2017) groundwater levels increased in half of the areas studied. In the later years of the conflict (2018-2020) groundwater decreased in seven of ten areas, with one area remaining constant. This was a new trend in four areas.

An early increase was relatively large in the southern Tihamah and the eastern highlands, moderate in the central highland plains and southern deltas, and minor in Ramlat. With the exception of the southern deltas, these were new trends and preceded a decrease in groundwater during the later years of the conflict (2018-2020). This indicates that the onset and current midpoint of the war represent change-points. These changes and their drivers are consistent with the narrative presented for western Yemen, albeit with some differences in magnitude.

Some regions do not follow the trend for western Yemen as a whole. In the northern highland plains there were stable groundwater levels in the early part of the conflict, before a drop-off in the later years. In the north west, western mountains and Abyan, there are continuous declines during the course of the conflict – all continuations of previous trends – and these areas suffer the greatest overall declines in groundwater levels. In the first half of the conflict Hajjah follows a similar trend. However, at the turn of 2017 this is reversed and levels have considerably recovered. With the data available, we have been unable to identify all the drivers for these trends. They are unlikely explained by rainfall or agricultural changes alone. Instead, we speculate they may be linked to conflict-induced population changes, for example drinking water needs locally or further afield. Further research and validation of these results is required.

Also incorporated into our analysis is a crude investigation into the relationship between precipitation and groundwater changes.⁴¹ A relationship could be defined for six of the areas between 2002 and 2010, and this was used to approximate the expected groundwater anomaly based on precipitation changes from 2011 onwards. All but one of these areas had lower groundwater than would be expected in the later years of the conflict, indicating that precipitation was having less of an influence on groundwater levels than it was prior to 2011 in the six areas where a relationship was defined.

Return to Contents

Table 2: Summary of key regional data. For each region the trend in groundwater levels is indicated with up (increase), flat (no change) and down (decrease) arrows, with a dot indicating if this is a new change. Where present, the correlation and monthly offset of the relationship between rainfall and groundwater is presented, alongside an indication of how the expected groundwater anomalies differ from the actual groundwater anomalies. Detail is also provided on the area (based on Copernicus Global Land Service maps) and type (Ministry of Agriculture and Irrigation statistics) of cropland. Information on the underlying aquifers manually estimated from the detailed overview of aquifer basins in Yemen by Jac van der Gun, Van der Gun Hydro-Consulting, 2015.

Return to Contents

2.3 Central highland plains, including Sana’a

The area we have named the central highland plains encompasses the higher elevations of the Sarawat mountain range and its flat fertile plains, and includes the capital Sana’a, the cities Dhamar and Rada and some of Ibb governorate. Of the areas studied, it contains both the largest population and area of cropland.

Figure 16. Key data plots for the central highland plains. Detail on data provenance in the Appendix.

Groundwater change

Groundwater started to recover in 2010, climbing to around the highest level on the satellite record in 2016. In 2017 it began to drop, to around the lowest level on the satellite record in 2020. This is in spite of higher than usual rainfall in the preceding three years, and the price of diesel following an upwards trend.

Early conflict

Groundwater recovery prior to 2014 is likely to be a result of social upheaval and political instability negatively affecting agriculture. With subsidies in place until 2014, the price of diesel did not fluctuate substantially, at least compared to later years.

When the diesel price shot up and the national grid collapsed in early 2015 the absolute and relative drop in nighttime lights was larger than any of the other regions. Although this is no surprise given it is the most urbanised region, the magnitude of the drop underlines the impact of the fuel crisis. Furthermore, there has been minimal recovery back to pre-conflict levels compared to, for example, the northern highland plains, where nightlights recovered to pre-conflict levels within a couple of years (see below). The difference in the nightlight data for the central and northern highland plains is an indicator of how the impacts of the conflict have been felt differently in different regions. In the central highland plains, groundwater levels rose in the months following the diesel price hike; in the northern highland plains they dropped.

Statistics provided by the MAI in Houthi-controlled Sana’a indicate that the area farmed for almost all crops decreased from 2013.⁴² Cereals, vegetables, pulses and cash crops show marked declines.⁴³ For some crops, the decline began earlier. Therefore, as groundwater was recovering, the crop volumes decreased, which would have had a negative impact on both livelihoods and food supply.

The one crop that did not decrease in area, but has remained steady since 2012, was qat. This region is one of the most significant qat-growing areas – in 2019 it produced 16% of all qat grown in Yemen, and 17% of the region’s total cropland was used for growing qat, according to the MAI statistics.

Later-conflict

It is plausible that the decrease in groundwater in the later years of the conflict is related to what Reuters reported as a “booming solar sector” in November 2019. This was said to be “transforming lives and energy sustainability”. The article highlights Dhamar as an area that has benefited from solar energy: According to a source at the local water authority, Dhamar’s water production had fallen to 30% of pre-war levels, but returned to 70-80% of pre-war capacity as a result of solar projects supported by international donors. “Our land had dried up but now it has come back to life thanks to the solar energy,” farmer Omar Homadi is quoted as saying. The central highland plains have the highest solar potential in all of Yemen.⁴⁴

Figure 17: Solar deployment in the central highland plains.

Figure 17 shows a new solar array that was constructed between Amran and Sana’a over a five-day period in March/April 2019. This illustrates the rapid turnaround in deployment and use of solar infrastructure. We understand the surrounding fields to be qat plantations.

The MAI agricultural statistics indicate that yields for a number of crops increased in 2019: cereals, vegetables, qat and fodder. For the latter three crops there was also an increase in area cultivated, although this was much less pronounced. An increase in yields is likely to be the result of more inputs, such as water.

Anecdotally there are reports of people in the Sarawat region turning to agriculture for their livelihoods as the conflict has caused urban household incomes to collapse. According to one Yemen water expert, as all the resources of society were directed to war, a combination of economic factors incentivised people in the region to return to the countryside, dig wells, intensify qat cultivation and expand the cultivation of vegetables.

This story of people moving (or returning) to the countryside is corroborated by another source, who told us: “Many old terraces have been reclaimed during this crisis using work for food programs and by the initiatives of Yemenis.”

Given that this region is the most populous area of the country, the humanitarian consequences of the water ‘running dry’ here would be the most catastrophic. The long-term historical depletion of the resource in the region, the recent accelerated depletion, and the region’s significance as a qat-growing area place it at significant risk.

Return to Contents

2.4 Southern Tihamah

The Tihamah is the arable Red Sea coastal plain of the Arabian Peninsula.  Perhaps Yemen’s most important agricultural region, it is sometimes referred to as the breadbasket of the country. Surface water drains from mountainous catchments to the east, over ground and underground through seven major wadis (valleys that contain water only in the rainy season) and is used in a seasonal flood-based spate irrigation system.

Groundwater change

From 2011 until mid-2018 there was a steady and marked increase in groundwater levels in the Tihamah study area. The peak in 2018 was unprecedented since satellite records began, with levels around 50 mm higher than the typical levels between 2002 and 2012, and the level expected based on the precipitation relationship. Although such a change may sound small, over the hundreds of square kilometres of each wadi aquifer it is a vast volume of water, likely experienced at some individual wells by changes in metres. Since 2018, however, groundwater has been steadily declining, despite an expected increase owing to rainfall.

Figure 18. Key data plots for Tihamah south.

Early conflict

The pre- and early-conflict groundwater recovery in the Tihamah is likely to result in part from increasing social and economic turmoil: 2011 marked the Yemeni revolution, the start of the Arab Spring and a significant turning point in Yemen’s recent turbulent history. The groundwater recovery, therefore, may signal the widespread collapse of livelihoods. At the point that Saudi Arabia and allies engaged in 2015, many people were already in a situation of extreme poverty. In June 2015, all of the Tihamah region was in the ‘critical’ or ‘emergency’ phases of acute food insecurity.⁴⁵

Rainfall was also above average in all years but one, and likely contributed to the recovery, however, the scale of the increase is larger than would be expected.⁴⁶ The worsening economic situation coupled with the escalating price of diesel would have made agricultural production unviable for many. According to Yemen water specialist Walid Saleh, “Many farmers abandoned farms as they couldn’t afford to pump the water.”⁴⁷

Between 2015 and 2019 an active front line moved around 250 km across the Tihamah, as the Houthis lost control of territory from the tip of the Arabian Peninsula to the Red Sea port of Hodeidah. Agricultural production in a war zone is clearly going to suffer, as indicated by CEOBS’ earlier research on agriculture in the Tihamah.

The conflict in the Tihamah had a three-fold impact on the labour force, which in turn negatively impacted agricultural production: the loss of life in agricultural communities; outward migration from the area; and a security situation that left many farmers unable to farm and labourers out of work. According to the latest available figures from the Displacement Tracking Matrix, net outward migration from the Al-Hodeidah governorate stood at 483,255 in November 2018.

A rapid assessment conducted in three governorates (including Al-Hodeidah) by the International Labor Organization (ILO) in 2015 found that in one year the agricultural labour force contracted by 50%. Even if such a figure is only roughly indicative, it demonstrates the social and economic upheaval going on during this period and its livelihood impacts.

Alongside the impact on the labour force, the impact of the war on agricultural infrastructure in the region has been significant. Note that up to 30% of all conflict incidents in a given quarter have occurred in the Tihamah study area.

Figure 18 shows that the area and yields of all crops have declined since 2010. The exception to this rule is qat, the growing area for which has remained an unlikely constant. This may indicate, however, that all of the viable land for qat is used for that purpose: the crop is better suited to the highlands and the Tihamah is a minor qat growing area, producing approximately only 1% of western Yemen’s qat in 2019.⁴⁸

A fraction of the land area that had been used to grow food in 2010 was being used to grow food in 2018. The volumes of fodder being grown were of a magnitude smaller, which is indicative of people selling livestock as their coping strategies were exhausted, or vice versa, livestock rearing becoming less viable as production of fodder decreased and prices increased.

Later-conflict

In 2019 and 2020 there were very significant drops in groundwater, of approximately 30 mm in total. The MAI statistics show that in 2019 there was a resurgence in agricultural production, with yields increasing for all crops. Pulses in particular increased notably in terms of the area on which they were grown, which was more than 50% greater than at any other time since 2007. This perhaps reflects the dietary needs of a starving population. Agricultural statistics for 2020 were not available at the time of writing (March 2021). Note that in 2019 there was above average rainfall, and in 2020 this greatly increased.

As noted previously, crop yield increases could be a result of environmental factors such as increasing rainfall, or agricultural intensification bought about through increased use of inputs such as groundwater. Coupled with the data on declining groundwater from 2018, it seems likely that solar-powered groundwater extraction is having an impact.

Visual analysis of satellite imagery indicates that solar panels began to proliferate in the region in 2018. The pattern, as shown in the carousel in Figure 19, appears to be for many small panels, as opposed to one big array. This may indicate groundwater extraction is taking place on a smaller, private scale, however, systematic research into the types of solar deployment has not been conducted.

Figure 19: Solar deployment in the southern Tihamah.

Salination of groundwater

Historically, the Tihamah region has suffered from saline groundwater as a result of seawater intrusion into coastal aquifers. A drop in groundwater levels migrates sea water intrusion further inland, resulting in the contamination of more freshwater resources with salt. According to Arvind Kumar of the UNDP, salinity in the Tihamah has doubled in one year.

Return to Contents

2.5 Northern highland plains

The area in this study defined as the northern highland plains stretches from the Saudi border in the north as far south as Marib, and includes the Sa’da governorate, notable as the homeland of the Houthis.

Figure 20. Key data plots for the northern highland plains.

Groundwater trend

Groundwater remained stable until 2017, when it began to drop substantially. In Autumn 2020, levels were far below those previously recorded.

Drivers

The pattern of decreasing agricultural production during the conflict is apparent in this region as in others; in 2012 the area under cultivation began to decline, though it is different for different crops. The period 2017-2018 is where they are at their lowest for most crops. Again, qat bucks the trend, maintaining a fairly constant area from 2012 onwards. In 2019, this area was responsible for approximately 10% of qat production in western Yemen. The crop yield results show less of a pattern, and more fluctuations, up until 2019, when all yields increase.

The 2015 drop in nightlights was relatively small compared to elsewhere in Yemen. They quickly rebounded and overtook pre-conflict levels by 2017. This may indicate a lesser impact from the 2015 diesel price inflation and crash, and go some-way to explaining the rapid fall of groundwater levels.

Visual analysis of satellite imagery around the city of Sa’da shows solar panels appearing in areas of agricultural production from around 2016. Agricultural production in the region now appears fairly advanced, with polytunnels commonplace. In some areas a resurgence of agricultural production is apparent, with previously barren-looking fields now green and solar panels in evidence. In other areas, desert has been transformed into cropland through the use of solar, as shown in Figure 21.

Figure 21: Solar deployment in the northern highland plains. The 2014 image may at first appearance look to be cloudy – it is in fact the bare desert which has been converted into agriculture by 2018.

Return to Contents

Part 3. Taking stock

3.1 Implications of falling groundwater

We have identified accelerated depletion of groundwater in Yemen since 2018, with a likely cause being a growing use of solar pumping. With extraction rates that were already regarded as unsustainable, such a trend raises the prospect of aquifers running dry or sinking to below accessible levels. This grave outcome and the path towards it pose humanitarian problems in relation to drinking and hygiene water, and livelihood loss as agricultural irrigation is constrained. Given the precarity of Yemen and past precedents, this may result in yet more insecurity and conflict. In some locations desertification will occur – natural vegetation will die, to the detriment of dependent ecosystems and populations, as well as local climate regulation.

With the currently available information we cannot forecast when thresholds for insufficient groundwater resource will be passed, although it is likely to differ spatially across the country – for example, coastal locations will suffer first as aquifers become increasingly saline. The consequences are also likely to be felt more quickly by the poorest, as it becomes more expensive to access the remaining water resources.

Secondary problems associated with aquifer depletion may also start to become more severe. In particular, land subsidence is likely to become a bigger issue – permanently damaging aquifers and infrastructure above ground. Relatively little research has been conducted in Yemen, but remote sensing studies have indicated that this was a significant problem near Sana’a. The country already has a high landslide hazard risk, and in some locations, this risk will be exacerbated by falling groundwater levels.

Return to Contents

3.2 Climate change

Climate change and its associated manifestations will be an increasing risk multiplier for Yemen. The country ranks as the 13^th most vulnerable to climate change, in particular via drought and extreme flooding. Already, the signs of climate change are visible in rising temperatures and decreasing humidity. The most recent global climate models suggest that Yemen will be warmer, with temperatures in the Arabian Peninsula set to rise by 1.8–2.7° C in the near future (2030-2059) under a business as usual scenario. At the same time, an increase in annual precipitation is predicted.

Information on expected climate extremes and geographical heterogeneity remain unknown but are likely to be significant – already CEOBS has shown spatially disparate precipitation trends between 1981-2019. Global climate models have a mixed capability at capturing such trends, and there have been no high-resolution Yemen-focused climate assessments from regional models. This need remains, especially given Yemen’s varying topography and climatic zones, and results from such studies in Saudi Arabia, which reach sombre conclusions about the habitability of some areas.

The role of dust in the climate and solar story is also important – it affects regional circulation and complicates climate predictions. There is a highly dynamic trend in dust or sand storms over the past two decades, and it modulates the efficiency of solar cells.

Return to Contents

3.3 Characterising the groundwater of Yemen

Prior to the escalation of war in Yemen, groundwater monitoring was being done at individual wells. How groundwater extraction rates translate into depth changes in wells depends on the local aquifer hydrogeology. Typically, well depth changes are expressed in metres: the country’s most recent report to the Convention on Biological Diversity states that water tables are lowering by between 2 and 6 metres annually in some basins.

In contrast, groundwater monitoring from space involves measurements in Earth’s gravitational field and soil moisture, which are combined to estimate groundwater change over a large geographic area. Data for smaller areas or individual wells is not possible – for Yemen, the spatial resolution of the data (the area over which the changes are measured) is in ‘pixels’ which represent approximately 50 km².⁴⁹

Such an area may incorporate various distinct aquifers, each with differing groundwater levels, and some areas where there is no groundwater storage. As such, signals are averaged out and the groundwater changes are reported in millimetres rather than metres. Nonetheless, over the area of land area being assessed, this may represent very large volumes of water.

Understanding the differences and limitations of these two approaches and results is important in terms of understanding their potential practical applications. Satellite remote sensing is useful in assessing changes at regional, national and supra-national levels, while other techniques need to be used to assess changes to the resource at the local – town or village – level. For the development of integrated and comprehensive water management plans, a combined approach is needed.

According to Walid Saleh: “There is a wider need for water managers, scientists, and policymakers to adopt new approaches to translate the large-scale water resources assessment enabled by satellite remote sensing of water resources into more effective Integrated Water Resources Management strategies at local scales.”⁵⁰

However, it should be noted that there are constraints on data collection in the north of the country; one source told us that it was difficult to conduct assessments there.

How much remains?

Both the remote sensing data presented in this research and the data formerly generated by monitoring individual wells show changes to the groundwater levels. The big question about Yemen’s non-renewable groundwater resources – how much there is left – remains unanswered. And for the aquifers that can be recharged, what would be a sustainable usage equilibrium?

Again, Walid Saleh: “As valuable as [the satellite] groundwater data is, it has a crucial gap: it cannot show the total volume of groundwater available in the aquifers it tracks, only the rate of decline. Without knowing when the aquifers will go dry, or when water tables will sink so low that they are effectively inaccessible, users and water managers are blind to the scope and severity of their problems.”⁵¹

However, Saleh notes that the drawdown of the groundwater table can give a good indication of the state of the aquifers if combined with water quality testing.

By combining results from satellite remote sensing, new field measurements and detailed knowledge of the hydrogeology, it may be possible to address the question of absolute water reserves in the future with sophisticated modelling. This would be a lengthy process involving a range of physical and geophysical survey techniques that may take up to a year, even for one location.

Given the heavily depleted groundwater in Yemen, its unregulated extraction, and the apparent arrival of a new technology driving accelerated extraction rates (solar power) there is a strong case for such work to be conducted. However, the timescales associated with this work, even if it were currently logistically possible, may be too long and too late. As such the imperative for more immediate implementation of sustainable groundwater management remains.

Return to Contents

3.4 Electronic waste

Another problematic environmental consequence of the deployment of solar power systems is the generation of electronic waste. Most manufacturers claim their solar panels will last 25 years, but many will be decommissioned early due to damage, defects and replacements. Commentators have noted that second hand solar systems are being sold with low quality control by recyclers in the United States to countries in the Global South with no electronic waste regulation.

Yemen already has a waste management problem, and solar power equipment is complex to recycle and contains toxic materials. Furthermore, it is often used in conjunction with batteries (though these are not required for water pumps) which are also hazardous waste. If such equipment goes to landfill – either formal or, as is more likely the case in Yemen, it is informally dumped – materials such as lead can leach out and create new environmental hazards. This could pose a threat to groundwater resources in the future.

Return to Contents

3.5 Solar stakeholders

Solar-powered groundwater extraction in Yemen is being deployed by the public sector, the private sector and humanitarian and development agencies. In order for the resource to be monitored and sustainably managed, there needs to be a multi-sectoral and multi-level knowledge sharing and co-ordination, to the extent that this is possible. In the following section we consider the relevant stakeholders, beginning with a review of their operational context, and present some case studies.

The state, authorities and local water governance

Fragmentation

Yemen is a divided country, governed by two parties at war with each other. The Houthis are the de facto government in what is usually described as the north, though this aligns more closely with the western Yemen study area as defined in this research. This is by far the most populous area, and includes the capital Sana’a and many other major cities. In what is often described as the south, but which includes most of the eastern, sparsely populated area of the country, the internationally recognised government is nominally in control. This has a seat in Aden, heavy support from Saudi Arabia and legitimacy afforded by the international community.

The truth is more complex than this north/south, Houthi/government oversimplification: the country has fragmented along many different lines and there are numerous active fronts at any one time. Nevertheless, we adopt this designation below.

There are dual systems of government in the two areas, with often two parallel government ministries. Before 2015 all state agencies were based in Sana’a. According to reports, most have remained there and are still functional, despite almost four years of the sporadic payment of civil service salaries. The Central Bank of Yemen is split, with headquarters in both Aden and Sana’a, and as of January 2020, two different bank notes. There are also two ‘state’ news agencies.

Water regulation

Prior to the conflict, the National Water Resources Authority (NWRA) had a mandate to manage the nation’s water resources. From its website it seems that it was once possible to extract data gathered from water well monitoring stations there. It would seem that the NWRA is still functioning in the north; according to Arvind Kumar, it is a policy-making body and as such plays a key role. A 2019 UK Department for International Development (DfID)-commissioned study reported that in north Yemen it was “still issuing licences and… cooperating with local councils to regulate water use”, however, this could not be verified by its authors. It is not clear whether there exists a parallel NWRA in the south: we have not established the existence of such.

The General Authority for Rural Water Supply and Sanitation Projects (GARWSP) is the implementation body. Kumar indicates that the GARWSP links with Water User Associations (WUAs), and that, through these, water is being tariffed. According to one source, the management of water resources is very active and functional at a governorate level.

There are contrasting reports regarding the level of control exercised over drilling rigs. One source stated that there is a lack of supervision, which has encouraged farmers to dig wells in the places they wanted, and resulted in a proliferation of wells. It may be that fines are collected in such situations. Another source stated that applying for a licence to drill a borehole is a tedious and bureaucratic process, therefore there is unlikely to have been an escalation in borehole drilling in recent years. We have also been informed that, in the north and south alike, the agencies are different but the bureaucracy the same.

A crippling aspect of the conflict in Yemen has been the sporadic payment of salaries of government employees, which have “become a playing card in the war”. It is hard to overstate the impacts of this in terms of the governance and rehabilitation of natural and other resources.

Water User Associations

At a local level, management of water resources is often in the hands of Water User Associations (WUAs).⁵² These are made up of elected community members and are recognised by, and under the authority of, the state.⁵³ The state does not recognise other models, such as those developed through the self-organising power of local communities.

There are varying reports in the literature regarding the effectiveness and impact of WUAs. They are often the conduit for interventions by external agencies, and as such presented positively. In reference to the Tihamah wadis,⁵⁴ Professor Martha Mundy says:⁵⁵ “WUAs have been created to devolve the problems caused by replacing traditional irrigation systems with high-tech, high-maintenance systems designed to privilege up-stream large landowners (“entrepreneurs” for the export market) and which have suffered over the years from siltation”.⁵⁶

According to a study on the Tihamah, published in 2018:

“The WUA itself is a well-recognised unit of local governance with a high level of acceptance in the study area by community members; however, many believed that the WUA, in terms of its ability to represent community needs, required improvement. In most cases, the WUA, in addition to dealing with well management, discusses access to water, and rules surrounding a water source, but did not facilitate forward planning around water resources. In one situation, the head of a WUA was a powerful private well owner who benefited from irrigation through his own farming operations. This highlights the potential conflict of interest that can occur when WUAs lack an accountable governance structure.”

The study raised the prospect that, when WUAs are supported by donor programmes, their life-span may be limited to a period of financial assistance.

Meanwhile, a publication by the Hague Institute for Global Justice, found that WUAs were regarded by the Tihamah Development Authority as a threat to local security, becoming a mechanism through which funds were captured and even weapons purchased.

WUAs intersect with existing traditional power structures, involving tribal or religious systems of governance and conflict resolution (see below). These decision-making bodies may all have varying degrees of control over how water resources are managed at a local level, although traditional methods of local conflict resolution have been disrupted by the national conflict. Regardless, perhaps the most effective method of control over water sources is financial – that is, through ownership.

A decline in the availability of water would put any system of governance under stress. Regarding the options to resolve some of Yemen’s water conflicts, the ELI’s Bruch observes: “It’s particularly difficult when water resources are shrinking. Conflict can aggravate this resource degradation, but it is important to recognize that even when conflict ceases there can be sustained reductions in water quality and quantity.”⁵⁷

Traditional watershed management

Yemen’s traditional watershed management evolved over thousands of years in the harsh terrain of the Arabian Peninsula. It is described by Anthony Milroy as the country’s “one greatest strength… the ancient, highly evolved, extremely resilient and eternally relevant, non-formal network, so deeply embedded across every catchment and sub-catchment throughout the country that it still quietly functions well below the radar and noise of government and development.”⁵⁸ It is a system which, he claims, has been “comprehensively and fundamentally ignored” by professionals ranging from politicians to scientists and development specialists for the past five decades. This may help to explain its scant attention in the literature.

Milroy claims that the geographical delineation of the tribal, and sub-district, sheikhdoms correspond exactly to watershed boundaries and the wadi centres below. “Across any catchment a sheikh has responsibilities to, and can call on duties from, all households,” he explains. The sheikh is supported by an Amal, or manager/land agent, and an Aqil, or water master, who has responsibilities and powers to apply all the local common laws pertaining to all local water management issues. This may include the timing of irrigation; resolving land and water disputes; and terracing, spring, spate and surface water run-off management. Reference is made to formal, written lists of agreed times, dates, and ownership names. The Aqil is elected at the tribal level and “if he is not up to the job he is politely dropped,” says Milroy.

Underpinning this system is an obligation and benefits-based rubric that requires collective action in the face of need. It may cover anything from raising money for education of the young, or fixing the house of a widowed mother, to “fixing a gravel sediment episode from a flash flood in the wadi,” and “repairing water cisterns, wells, spate water offtakes, and terracing collapses that affect households without males and therefore neighbouring lands,” says Milroy.

Yemen’s population increased from 6 million in 1970 to 30.3 million in 2021. This, coupled with the abandonment of rainwater harvesting systems, dwindling groundwater, periods of violent conflict and worsening poverty, would have placed additional strain on traditional water management systems. Nevertheless, Milroy claims that they “remain fundamental to the very fabric of the landscape and the society at a rural level.”

According to another source, the adaptive traditional management system has been “weakened considerably during the last four to five decades of outside influence by the government and development agencies.” Further, the source advised that it is relevant to consider the application of Islamic (Sharia) law, a fundamental principle of which is that it is not allowed to cause harm to yourself or to others by your actions. Through generating a dialogue about respect for this tenet of Islamic law, it may be possible to reach the hearts and souls of Yemenis themselves in such a way that Yemen’s groundwater resources – and by extension livelihoods, environment and society – can be saved.

In-country humanitarian and development agencies

General operational context

Aid agencies in Yemen work within the constraints of a very difficult operating environment. In the north, the Houthis have blocked access to the population and tried to enforce agreements that would make it very difficult for them to maintain their humanitarian principles. This has also had a silencing impact on aid agencies and workers, as people are scared to speak out for fear of losing their permission to be there. The Houthis have also diverted aid away from intended beneficiaries, leading to repeated stand-offs with aid agencies.

Aid obstruction is also reported in the government-held south and east of the country. Saudi Arabia works closely with the internationally-recognised government and has given vast contributions to the international aid effort.

There have been allegations of corruption at UN agencies.⁵⁹ Added to all this, both humanitarian and development agencies face chronic underfunding and are operating in the conditions of a pandemic.

The international community plays a threefold part in Yemen’s conflict: It funds and operationalises the aid programme, which is the world’s second-largest after Syria, and the majority of this is administered through UN agencies. It has a key role in peace negotiations such as the Stockholm agreement, which was endorsed by the Security Council in 2018. Thirdly, it supplies the Saudi-led coalition with arms, with 96% of Saudi’s weapons in 2019 being supplied by the five permanent members of the UN Security Council.⁶⁰

The paradoxical nature of these roles was not an issue for World Food Programme Executive Director David Beasley who, in advance of the 2021 High-Level Pledging Event for the Humanitarian Crisis in Yemen, said: “To me those that are engaged in the conflict should be the ones paying for it.”⁶¹

Groundwater in humanitarian and development assistance

Multiple international agencies have provided solar pumps for drinking water wells in Yemen. Some have also provided solar pumps for irrigation purposes, including the FAO, International Organization for Migration, and UNDP.

The World Bank supervises the Smallholder Agricultural Production Restoration and Enhancement Project for Yemen, a $36 million agricultural project that is being implemented in-country by the FAO and the Yemen Social Development Fund. Senior Agribusiness Specialist and Task Team Leader Rufiz Vakhid Chirag-Zade told CEOBS that, whilst the project supports the rehabilitation of community water infrastructure, there are no solar power elements involved.⁶² This is because “the project activities do not require deployment of solar power.”⁶³

We spoke with UNDP Project Manager Arvind Kumar and Abdu Mohammed Seid, a water engineer for the NRC, and found that there was not a consensus in-country regarding the impacts of the solarisation of groundwater abstraction.

It is possible that the contrasting perceptions regarding solar pumps are a result of the different perspectives of humanitarian and development organisations. Whilst humanitarian agencies such as the NRC primarily intervene in order to provide drinking and hygiene water, the interventions of the UNDP focus on livelihood security and agriculture, and water for the purposes of irrigation. The quantities involved in extracting for these two needs are very different. Experience of different locations will also inform their perceptions: for example, Kumar works with farmers in the Tihamah, where the issue of groundwater salinity is becoming increasingly apparent.

WASH clusters

Co-ordination amongst agencies in Yemen largely happens through the national and provincial Water, Sanitation and Hygiene (WASH) clusters. These are formed of multiple agencies “with the objective to coordinate the WASH humanitarian response in Yemen”. Their core activities include supporting service delivery, planning and developing strategies, and addressing accountability to affected populations. In the north, the GARWSP is part of the wash cluster.

According to Seid, to some extent there is knowledge sharing within the WASH clusters about groundwater resources. He stated that generally there is an understanding that groundwater is depleting 5-10 cm a year, as a result of a study that was conducted by the Ministry of Water and Environment for the internationally recognised government in Aden.

Saudi development assistance

The Saudi Development and Reconstruction Program for Yemen (SDRPY) provides both humanitarian and development assistance in Yemen. It has initiated more than 188 projects including in the energy and water sectors. As of February 2020 it had “rehabilitated and drilled over 40 wells in Yemen, equipping them with solar powered pumps to supply enough water to meet the daily demands of local communities for agricultural self-sufficiency.”

Figure 22: Locations and types of Saudi Development and Reconstruction Program for Yemen projects, as disclosed by @SaudiDRPY (Twitter).⁶⁴

Saudi Arabia’s financial support for the UN’s aid missions in Yemen dropped by 65% from $1.3 billion in 2019 to less than $500,000 in 2020, making a massive difference to the overall sums received, with ramifications for all UN agencies. The same year it increased the profile of its bilateral aid provision, with an apparent focus on development, as a pose to humanitarian, aid.

In December 2020, at a meeting between the SDRPY and the Islamic Development Bank, it was agreed that they would set up a “unified and comprehensive development strategy” with the Yemeni Government and “all international donors”. Future projects would include those focussed on “enhancing the country’s agricultural exports.”

Yemeni exports to Saudi Arabia stopped in 2016 but in 2019 restarted with a small volume of fruits and vegetables. Given the historical importance of Saudi Arabia as an export market for Yemen’s cash crops and the over-exploitation of its own groundwater reserves, this move into Yemeni agriculture – albeit under the guise of development – could present a further threat to Yemeni groundwater reserves. There is no indication that the SDPRY has policies in place, or uses mechanisms at the pumps, to prevent over-exploitation of groundwater.

Figure 23: A solar water project installed by the SDRPY. Photo credit: Saudi Press Agency.

In September 2020 a Saudi water project in Aden was reported to benefit 1.5 million beneficiaries. No mention was made as to whether any of the wells would provide water for agricultural purposes. However, at a February 2020 meeting between Saudi officials and the then-UK Foreign Office and International Development minister Andrew Murrison,⁶⁵ the SDRPY supervisor claimed there was “close cooperation with international development organizations,” and that “existing projects in the Agriculture sector that require Solar Energy have benefited greatly from this approach.”

The SDPRY official also said that the programme intended to expand its reach across all Yemeni governates with the opening of several offices. There are political implications of such a statement given that many governorates are under the control of the Houthis.

There is high-level engagement between Saudi development officials, UN agencies, and bilateral donors,⁶⁶ though no links were found with organisations operating on the ground.

Yemeni civil society

There are numerous active Civil Society Organisations (CSOs) in Yemen: in 2016 approximately 12,000 associations and foundations were legally registered. Throughout the conflict they have suffered multiple violations, from direct physical abuses against personnel, to restrictions on their freedom to operate, to defamation in social media. Nevertheless, Yemeni organisations continue to play a role in delivering aid and human rights services,⁶⁷ and in international advocacy, advancing the key issue of women’s meaningful participation in the peace process. The war has increased the need for self-help within communities.

The private sector

The private sector has a crucial role in Yemen and its future development, particularly in light of the inadequacy of state provision and regulation, and the current and probable future underfunding of aid agencies.

Solar sector

In the solar sector there are signs that, in conjunction with international agencies and local NGOs, the private sector can play a role in addressing the regulatory void that exists in the absence of a fully functioning government.

The surge in demand for solar energy after 2015 led many entrepreneurs to supply solar equipment. But with regulation absent it became apparent that the required checks on suppliers and equipment were threatening the development of the sector. In response, the solar energy platform the Green Pages was launched by a mix of public, NGO and private actors in 2016. Its activities included awareness raising, information dissemination and qualitative assessments of products.

The Green Pages’ website is currently down, however, and CEOBS has been unable to establish the current status of the project.⁶⁹

Water sector

The vast majority of water resources in Yemen are privately owned: some 90% according to the UNDP’s Arvind Kumar. It is hard to overstate the significance of this in terms of understanding Yemen’s water dynamics. Water is not considered a public good but a commodity with a huge value chain, says Kumar. Wells are a source of income as well as water for their owners, who sell it to tankers for distribution elsewhere. Kumar asserts that there is little watershed management. Water is cash-generating, and there are no incentives to manage the extraction of water for sustainability purposes.

Changes in the how the private sector operate can have impacts on groundwater. According to one source, drilling rig owners now accept payments in instalments after they have drilled a well.

Return to Contents

Part 4. Interventions

Water resource management is a key element of post-conflict peacebuilding, as has been well-documented elsewhere. For Yemen, the post-conflict context does not apply, nor does it seem a likely prospect in the near future. Environmental peacebuilding initiatives such as the employment of ex-combatants in the rebuilding of environmental infrastructure are distant aspirations. But although the situation is complex and the conflict dynamics present huge and ever-changing challenges, there is an urgent need for action on multiple levels to assess and address the issue of groundwater decline.

In February 2021 we asked the Prime Minister of the internationally recognised Government of Yemen Maeen Abdulmalik Saeed what was being done to prevent the unsustainable extraction of groundwater as solar power is being deployed for agriculture in Yemen. He acknowledged that “… there is no regulation on the part of local authorities of the use of solar powered water pumps, and this is one of the risks in a country that faces water security issues like Yemen.” He said that the government was working on the local authority level to ensure that the system does not fail.⁷²

In order to avert the humanitarian and environmental consequences of the exhaustion of Yemen’s groundwater, a multi-sectoral approach is required, encompassing technical, social, economic and political components. Action is needed at local, national and international scales, however, the potential for intervention is seriously compromised by the realities of the conflict.

In Anthony Milroy’s words, the “failed state, catastrophic starvation, and complete breakdown of all societal norms” means that any recommendations that are made must be heavily caveated. In the summary of recommendations table we address this by applying a temporal framework, that is, we identify what is possible today and what should be addressed in a medium or long-term time frame. None of the interventions recommended are for a post-conflict period as Yemen’s future is so uncertain. However, we have also not excluded interventions that may appear to be impractical at the present time. This is because the security conditions vary across the country, and because stakeholders may nevertheless be able to take steps towards implementing them in part.

One issue to raise is that due to the breakdown of governance and fragmentation of the Yemeni state, top-down interventions are potentially less likely to work as there may not be the structures in place to implement them. They are also less likely to be either empowering or egalitarian. Identifying ways to work on the local level are therefore vital.

Due consideration needs to be made of Yemen’s local knowledge and traditional water management structures and systems, into which groundwater monitoring and education could be folded.

Return to Contents

4.1 Groundwater monitoring

Three possible methods for groundwater status monitoring are remote sensing, geophysical surveys and depth measurements taken at wells. A comprehensive national assessment would involve all three of these; in Yemen the current prospects of this are remote.

This research presents preliminary earth observations that show significant groundwater changes from 2018 onwards, and there is a clear need for continued monitoring by specialists. This will act to verify our results and reduce the significant uncertainty, as better processing techniques are developed and new ancillary data sources come online.

We recommend that an expert roundtable is convened, featuring domestic and international practitioners, including scientists and engineers. The group’s initial task should be a reanalysis of existing hydrogeological and groundwater data, with the goal of producing a technical ‘state of the groundwater’ report. This document would also include a roadmap indicating what new data is required, and what technical solutions may be available in the near, medium and long-term.

Likewise, we recommended that the potential for the resumption of well-based groundwater monitoring is assessed by national and international agencies, regional and national authorities, and local communities. To enhance local understanding of aquifer recharge rates, this should be combined with precipitation data and existing hydrogeological information.

Where traditional water management structures, Water User Associations or other water user groups or civil society organisations are functional, there could be a role for citizen science to play in groundwater and precipitation monitoring systems.⁷³ This may help engender a wider sense of stewardship of groundwater resources.

Within organisations that operate or contribute to the maintenance of multiple wells, systematic monitoring of groundwater levels should be implemented. There may be a role for international institutions or civil society actors to assist in the development of context-appropriate technologies. Private well owners may be incentivised to monitor groundwater levels through financial or other means, for example, technical or maintenance support.

Further research should also be conducted on the secondary effects of falling groundwater. In particular the land subsidence risk across western Yemen, using recently available satellite Interferometric synthetic aperture radar data.

Assessing the deployment of solar energy

A comprehensive assessment into the scale of solar deployment is required to help understand the locked-in and future risk of groundwater depletion and hazardous waste disposal from solar pumping. As noted, there is an absence of official data. We recommend a complementary approach where local on-the-ground surveys can help train and verify country-scale estimates based on a machine-learning algorithm identifying solar panels using high-resolution satellite imagery. These surveys could be undertaken by a range of actors: development agencies, civil society or governmental agencies. The size of arrays could potentially provide a qualitative indication of groundwater depth, if assumptions can be made about the number or depth of associated boreholes.

Groundwater analysis to be part of the solar discourse

Ultimately, consideration of groundwater levels must become part of the discourse around solar-powered groundwater extraction in Yemen.

Return to Contents

4.2 Knowledge exchange

In-country

Should well-based groundwater monitoring be adopted at scale, the question arises as to which organisation or authority would collate and share such information, and to what purpose.

Programming that fosters learning and knowledge exchange is identified as a key component of adaptive post-conflict water governance and management.⁷⁴ In Yemen, the war between the different administrations that are in control – or nominally so – in the different areas of the country currently renders groundwater knowledge exchange difficult on a national level. Furthermore, many parties to the conflict have crossed humanitarian red lines, and demonstrated a willingness to use water, food, electricity and access to the population as targets or bargaining chips in the war. The generation of knowledge about groundwater resources may in itself create risks that it may be used against local populations or parties to the conflict.

Nevertheless, Yemen’s people and water resources cannot wait for the war to end before action is taken to protect the resource and people’s access to it. There is a need for more knowledge sharing amongst international agencies working in-country. The divergent views on the impacts of solar-powered groundwater abstraction at the UNDP and NRC identified during this research is an indicator of this need. It is recommended that this report is circulated to all stakeholders.

An appropriate forum for knowledge exchange between agencies may be the existing WASH clusters, which could expand their remit to include groundwater monitoring at wells. Such information could be used to inform the programming of interventions, for example, the suitability of purchasing water from particular sources to be supplied to populations in other areas. As far as we have been able to establish, at present the source of trucked water is determined purely by logistics and market mechanisms; the sustainability of the source is not factored in.

A Yemen Knowledge Hub to connect with the international community

Given the need for knowledge exchange about Yemen’s groundwater resources, and the deployment of technology suitable for the Yemeni context, it is recommended that a feasibility study is conducted into establishing a hub for knowledge exchange and technology transfer between Yemeni civil society and international actors. These actors may be international agencies, academia, the private sector or civil society organisations. Particular effort should be made to engage with, support and include Yemeni academics studying overseas. The objectives would be to connect Yemeni civil society and the agencies that support it with the international community in order to enhance each other’s work, support sustainable interventions and respond to needs or requests from organisations working in Yemen. Vice versa, local surveys and citizen science are needed to validate remote sensing results and could feed into groundwater monitoring at scale.

Examples of appropriate technologies that are needed for Yemen include: tools for monitoring groundwater levels and quality, and methods of reducing manual labour in the reconstruction of terraces (see below).

Return to Contents

4.3 Sustainable livelihoods

Water rights and access

Whilst recognising that all interventions need to take into account the existing water management structures, there is a general country-wide need to address water rights and access across watersheds using participatory frameworks.

For example, in the Tihamah, groundwater salinity along the coast is increasing as a result of extraction by farmers in the lower catchments. The root cause of this problem lies in inequitable water distribution with farmers in upper catchment areas. The deployment of technologies alone – either solar or desalination – will not solve this.

Enhancing participatory groundwater management may involve existing tribal or religious decision-making processes, as well as Basin Committees, Water User Associations and Water User Committees. But it should also include those not represented in these existing structures, for example, women and marginalised communities, and involve both upstream and downstream waters users.

Reviving Yemen’s terraces and traditional water harvesting techniques

The value of Yemen’s traditional systems of water and land management must be recognised by the international community and organisations working in Yemen. Yemen’s ancient terraces utilise and can equitably distribute rainwater, and may also aid aquifer recharge. Conversely, the abandonment of terraces has increased surface water run-off and exacerbated flooding downstream.

At the time of writing, famine is spreading across Yemen at a rate not previously experienced. Yemenis need to be supported in growing their own food, in a sustainable way. A sustainable system tailored to the Yemeni environment was developed over millennia, and can be repaired and reinstated now, with international support. Traditional approaches should be a key component of future groundwater management plans.

According to one source: “The only path to sound land and water management in Yemen is through revitalizing the age-old systems of terraced farming using modern technology to reduce manual labour and maintenance. In an environment like ours only time-tested practices are viable. But with modern technologies the burden on the population could be eased.”⁷⁵

It may be important to note that the climatic conditions under which future terrace restoration would happen are likely to be different from the conditions historically, with more evaporation and more rainfall. It’s unclear how this would impact the efficacy of the terraces.

Redesigning donor assistance

The UN’s aid delivery programme must be reviewed and restructured to be made more flexible and suitable for a more diverse group of actors, in particular smaller and women-led organisations. This would have multiple impacts:

• Empowering and supporting those already engaged with delivering water, addressing water needs and solving water conflicts at a very local level.
• Local economies would benefit if more money were channelled at a local level, particularly into isolated rural communities.
• Access to isolated rural communities in Yemen has become harder for international agencies, it is therefore imperative that local groups are sought to help to resolve water access issues in hard-to-reach territories.
• Smaller organisations can be more efficient, and a relatively small amount of money can have a lot of impact, for example, through the rehabilitation of a water resource.

Beyond solely funding the provision of water access improvement measures, projects should consider the proposals outlined here, namely: groundwater monitoring by local communities; the deployment of water saving technologies in agriculture and the provision of appropriate finance for small producers to enable the implementation of such technologies.

Water-saving technologies in agriculture

Whilst solar-powered water pumps may meet the immediate needs of farmers, they are not a panacea to Yemen’s water stress and in the long term may exacerbate it. Alongside the deployment of such systems for agricultural purposes it is imperative that water-saving technologies such as drip irrigation systems are also rolled out. As noted by the UNDP, such systems may require substantial behavioural change and are expensive for farmers.

In addition to the supply of equipment, to overcome these barriers training and financing schemes that are appropriate for small farmers and women are necessary. These could be combined with the generation and exchange of knowledge on groundwater resources and aquifer recharge rates at a local level, for example, through citizen science programmes.

Agricultural production using greenhouses also requires less water. Satellite imagery indicates that greenhouses are becoming more prevalent in some areas.

Threshold values at water pumps

There are ways in which groundwater over-extraction can be prevented or reduced at individual wells. At the design phase, the size of the pump is critical for preventing over-abstraction or conversely, the withdrawal of too little water to meet the needs for which the project was designed.

The ICRC has produced guidelines for test pumping at water wells, which may be useful in terms of determining a borehole’s efficiency and optimal production yield.

Systems to prevent over-extraction, so that yields do not surpass a threshold value, can also be installed. However, identifying the threshold value is a difficult task as it is dependent on a multitude of factors. If the threshold is chosen based on the aquifer properties then it requires knowledge of the basin hydrogeology, extraction activity elsewhere in the aquifer, and future rainfall patterns in relation to the recharge potential of the aquifer. Comparatively, if the threshold is set based on the minimum water requirements of the cropland, then this will vary based on the climate conditions in a given season, the efficiency of irrigation, and the relative role of rain-fed watering.

Toolbox on Solar Powered Irrigation Systems

The Toolbox on Solar Powered Irrigation Systems is a multi-agency product that addresses risks related to system efficiency, financial viability and the unsustainable use of water resources. It includes a Water Resource Management checklist for “systematically considering the sustainable and legitimate use of water resources”. This addresses the water source, including the type of well and hydraulic characteristics; ecological factors; efficient water management; water administration and access; and water extraction licenses.

It is recommended that all those deploying solar power irrigation systems in Yemen use this tool, if not already doing so.

Sustainable livelihood support

Where livelihoods in Yemen are dependent on unsustainable groundwater extraction, efforts should be made to support alternative livelihood development. According to Walid Saleh, some qat farmers have found a monetary advantage in switching to vegetable production where greenhouses have been introduced. Such interventions may be appropriate in areas where qat farming is having a detrimental impact on water tables.

Where well or borehole owners extract water for monetary purposes, there may be interventions that could promote alternative revenue-generating activities.

Environmental impact assessments

All large-scale well and borehole operators should undertake and publish Environmental Impact Assessments that address potential groundwater impacts before drilling new boreholes.

Return to Contents

4.4 Environmental peacebuilding

The peace agreement

According to Bruch of the Environmental Law Institute, peace agreements have a strong role in setting the post-conflict recovery agenda: “If water and other environmental factors are not in the peace agreement often it’s difficult to get the resources and political attention for those issues in a post-conflict period.” Funders tend to focus on implementing the peace agreement, and justifying expenditure for water is harder if it is not in there, he says.

The inclusion of environmental provisions can be difficult to argue for when peace agreements are being formulated, however. “Do you want to delay having peace until we get these environmental provisions in place when people are dying every day?” asks Bruch. On the other hand, sometimes they are low politics and can be addressed constructively as part of a process of building confidence, according to Bruch.

Yemen’s future will depend on the sustainable management of its groundwater resources, so it is imperative that such considerations penetrate higher-level peace negotiations. “If you’re not able to address the environmental dimensions, the peacebuilding efforts will languish and there is a risk of a relapse to conflict,” says Bruch.

Compliance with International Law

Damage to water infrastructure in Yemen has exacerbated human suffering due to the spread of diseases and drinking water becoming harder to obtain. It also has implications for rebuilding; there will be limited – if any – investment in major water or energy infrastructure whilst the threat that it will be destroyed looms large.

International Humanitarian Law (IHL) is designed to “protect civilian objects, including those indispensable to the survival of the civilian population.” An October 2020 Group of Eminent Experts’ report to the UN Human Rights Council found “reasonable grounds to believe that the parties to the conflict in Yemen are responsible for pervasive and incessant international human rights law and IHL violations, many of which may amount to war crimes.”

The international community has a role to play in pressuring all sides of the conflict to uphold standards of international law, through diplomatic, economic and other means. Restricting the arms transfers that facilitate violations would be in line with the recommendations of the Group of Eminent Experts. In 2020 it reiterated its call “for third States to stop transferring arms to parties to the conflict… No State can now claim to be unaware of the scale of violations occurring in Yemen.” Civil society actors can also play a role in this process.⁷⁶

Return to Contents

4.5 Private sector

According to Luqman, international agencies and the donor community have not supported or recognised the importance of the private sector in Yemeni development. This has led to many wasted opportunities for collaboration that could lead to enhanced efficiency and sustainability with local actor partnerships and local community led initiatives.⁷⁷ Efforts to support nascent sectors such as renewable energy can come from a number of directions, including trade associations and journals.

The trading association GOGLA, the global association for the off-grid solar energy industry, supports its members through market intelligence, knowledge-sharing, advocacy and promoting industry standards and guidelines. GOGLA has two Yemeni members, including the Musanadah Foundation for Development, which was instrumental in setting up the Green Pages.

Civil society internationally can support the development of sustainable technologies that are appropriate for the Yemeni context, for example, technologies to monitor and limit groundwater extraction will be important for solarisation of the water sector in countries such as Yemen. It could also play a role in raising awareness where inappropriate technologies may be being deployed.

Solar equipment take-back strategies should be included at the tendering stage of major projects to reduce the generation of e-waste, an issue for companies and regulatory authorities alike.

Return to Contents

Conclusion

The humanitarian situation in Yemen is one in which vast numbers of people are in dire need of water and food. This has been created in part by a dependency on diesel for the supply of water, with knock-on impacts for agriculture, livelihoods, purchasing power and the current humanitarian situation. The availability of diesel is at the mercy of markets, and supply routes that have been weaponised and attacked during years of war. Solar power – which comes with no ongoing fuel costs, reduces CO₂ emissions and supports decentralised energy systems – provides a vital route out of this relationship.

But the deployment of solar power to extract groundwater resources comes with the risk of increased, unchecked and unsustainable extraction. The results of this research indicate that these impacts are already detectable from space.

Yemen’s people have no place to move to should the water run out. It is imperative that steps are taken to avert this situation. This report comes with a set of recommendations for all actors, from local water users to international agencies. However, efforts need to be made to tailor interventions to the local context. In the words of Anthony Milroy, “It is not enough to assume that technically correct proposals will find their application effective without firmly binding them to Yemen’s truly unique and relatively intact watershed management heritage.”

Experiences on the ground in Yemen have demonstrated that, even during this period of violent conflict, co-operation around water resource management can help to build co-operation and trust within and between communities, and spark the prospect of peace. “People are good,” the NRC’s Abdu Mohammed Seid told us. “People can support each other to some extent. It depends on the place, Sheikh leaders, discussion and community.”

With thanks to: Abdu Mohammed Seid, Anthony Milroy, Arvind Kumar, Carl Bruch, Professor Martha Mundy, Mohamed Nabil Hayel Saeed, Muna Luqman, Peter Salisbury, Taha Al-Washali, Walid Saleh and the others who spoke with us on the condition of anonymity. 

Return to Contents

Appendix

High resolution versions of figures in this report can be downloaded from this link. Data and code are available on reasonable request.

I. Data plots for each district groundwater region

Figure A1. Key data plots for areas with distinct groundwater trends. The provenance of data in these plots and data used throughout the report is provided below.

II. Data provenance

A. Diesel

Diesel data from the World Food Programme November 2020 food prices dataset, averaged to each region.

B. Nighttime lights

Nightime light data from the Visible Infrared Imaging Radiometer Suite (VIIRS) Stray Light Corrected Nighttime Composites V1, masked for cloud-cover during processing in Google Earth Engine.

C. Agricultural statistics

The data on agriculture are based on official statistics from the Sana’a based Ministry of Agriculture and Irrigation (MAI) yearbooks 2011-2019. This data is reported by governate and so to report on the groundwater regions used an ancillary dataset was required – Copernicus Global Land Service 100m Land Cover maps. Where the 2015 crop fraction was more than 15% we defined a pixel as cropland. The percent of cropland each governate contributed to a groundwater region was used as a weight to calculate a representative mean from the MAI data. As such, the values reported for each groundwater region are a processed metric that ought to only be considered approximate.

n.b. the MAI is based in Sana’a and under Houthi control, and therefore liable to political manipulation; this may not be the case here, but such a risk is one of the problems of using data from conflict parties

D. Groundwater change

As the well measurement network has collapsed during the conflict (2014-present), we turn to remote sensing to investigate changes in groundwater storage in Yemen. We seek to understand how these changes may drive or result from surface activities associated with the conflict, contextualised by the longer-term changes to groundwater storage.

Through satellite data we are able to observe changes to groundwater (GWA), once we account for the sum of all water storages at or below the land surface, i.e.  the terrestrial water storage anomaly (TWSA):

TWSA = GWA + SIEA + SWA + SMA

Where SIEA = snow and ice water equivalent anomaly, SWA = surface water anomaly (i.e. from lakes, reservoirs, rivers) and SMA = soil moisture anomaly. As Yemen has negligible snowfall and surface water* we only need TWSA and SMA data to calculate the groundwater change (GWA).

* The canals, dykes and terraces used in spate water management in the west of Yemen may represent a significant surface-water component that we are unable to control for. These features are not incorporated into any model simulations, nor can be explicitly observed given their small sizes (even from high resolution optical imagery). 

Terrestrial water storage anomaly

Terrestrial water storage anomaly data was acquired via gravity measurement techniques by NASA’s Gravity Recovery and Climate Experiment (GRACE) mission. GRACE spherical harmonic solutions are available from three different processing centres at JPL, CSR, GFZ. Although the differences are mostly very small, and lie within the error bounds of the GRACE solution itself, to best reduce uncertainty, the recommended solution was to use the average of the three solutions.

However, this approach has since been superseded by the JPL ‘mascon’ solution, which constrains the raw grace data to a geophysical model resulting in: greater spatial resolution allowing use in smaller basins, a separation of the signal from land and ocean, fewer leakage errors and smaller gain factors. This ‘mascon’ solution is the one we use in this analysis.

The mascon data includes errors and uncertainties resulting from the measurement and data processing steps. Gain factors were applied to the mascon data to compensate for signal removed during data processing, in the destriping and smoothing steps to remove correlated errors. Conservative uncertainties are provided for each 0.5 degree grid cell, although these represent the uncertainty of 3-degree native resolution of the mascon; the uncertainties for the 0.5 degree grid cell by itself would be higher if quantified.

Earthquakes can cause sufficient displacements of the lithosphere to generate a change in Earth’s gravity field which GRACE would measure. According to the https://www.seismicportal.eu/ database, there have been no earthquakes in Yemen over 6 Mw during the GRACE time period.

Soil moisture anomaly

Most studies using GRACE data to reveal the groundwater changes rely on data from the Global Land Data Assimilation System (GLDAS) for the soil moisture component. GLDAS has three main data components, 2.0, 2.1 and 2.2, each assimilating a different level and type of observational data into their simulations, although this does not include observational data on soil moisture. These simulations are performed by a range of models: Noah, Catchment, the Community Land Model, and the Variable Infiltration Capacity. Thus, for a given location there exists a range of soil moisture values from GLDAS – in some cases this range can be large and disparate. Furthermore, the data is retrospectively updated with each model run, so changes in the model version or other input data may lead to noticeable changes in the data.

Together, these drawbacks are a problem for Yemen, as the scale of the soil moisture changes are approximately the same as those of groundwater. Thus, more reliable soil moisture data is required than other regions of the globe where the groundwater signal dominates. Therefore, instead of using GLDAS like most prior studies, we have used observation based soil moisture data from the Copernicus Climate Change Service (C3S). The combined soil moisture product harmonises and merges observations from active and passive microwave sensors. This analysis utilises the provided “sm_uncertianty” parameter which represents the error variance of the harmonised data sets estimated through triple collocation (TC) analysis.

Differences between derived groundwater anomalies were negligible when using the C3S product and a similar product ESA CCI. This ESA dataset actually includes more microwave data (from the SMAP mission) but we did not use this in this analysis as the dataset did not cover into 2020 at the time of data processing. For the 2001 to Dec 31^st 2019 period data processed using the CS3 v201912.0.0 algorithm was used, and for data in 2020 the CS3 v201812 algorithm was used.

To the best of our knowledge this is the first study to use the C3S soil moisture data with GRACE data to retrieve groundwater storage. The reason for this is that we are not using any other terms in the terrestrial water equivalent budget, which necessitate the use of a hydrological model (such as GLDAS) in other studies.

Region identification

The hydrogeology of Yemen is complex with many different aquifer types and traps (Fig. A2), most of which are smaller than can be resolved by GRACE. Hence for the first iteration of this analyses we used coarser generalisation of the main basins using the World-wide Hydrogeological Mapping and Assessment Programme (WHYMAP) product, developed by UNESCO (Fig. A3). However, it became apparent that because of the complex hydrogeology, differences in geography, climate and surface drivers, this was an oversimplification too far – changes were too heterogenous within these basins.

Instead, the groundwater changes in Yemen were segregated into self-similar areas via k-means clustering of the groundwater anomaly time-series since the start of unrest in 2011. The optimum number of clusters was identified manually – by identifying the number which resulted in the lowest associated error in the calculated mean groundwater anomaly time-series. This approach is perhaps different to the standard approach of manually defining the local areas, for instance based on the aquifer characteristics or conflict dynamics, but in so doing avoids bias and ensures interesting trends are not missed.

Many of these regions of interest comprise areas smaller than the native GRACE 3° grid cells and thus there is a danger of overinterpreting the data. However, by using the k-means approach it is hoped the problem of “signal leakage” is minimised. Nonetheless, given this uncertainty, the results for our clustered regions can only be presented as indicative, pending more expert analyses.

Figure A2. Detailed overview of aquifer basins in Yemen. Source: Jac van der Gun, Van der Gun Hydro-Consulting, 2015. The map follows on from working in earlier publication “The Water Resources of Yemen”, 1995.

Figure A3. Major basins in Yemen from the WHYMAP project.

The data processing steps undertaken are summarised by the flow chart in figure A4. Uncertainty was propagated through the data processing using Monte Carlo simulations where the error was correlated, and standard error propagation rules where not. All analysis was undertaken in the MATLAB programming environment. The groundwater storage anomaly and precipitation anomalies are presented as time-series for each region of interest.

Figure A4. Flow chart of data processing steps.

E. Precipitation

Here we use precipitation estimates from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) is dataset. This product combines satellite precipitation and cloud data with in-situ station data to create gridded rainfall time-series, also accounting for biases introduced by complex terrain. We use the daily 0.25° version of the data, summed up to the GRACE time base and latitude/longitude grid.

The blending algorithm used to produce the CHIRPS data does not include uncertainty information; this represents a substantial weakness of the product and means we cannot present uncertainties associated with the precipitation anomalies calculated. Apparently, work is ongoing to provide standard error fields, which may be available in future CHIRPS releases. In nearby Ethiopia, which has a somewhat similar climate, Dinku., et al (2018) found CHIRPS had high skill when compared to independent rainfall assessments.

It must be noted that the ground gauge data going into CHIRPS from Yemen ceased after the start of the conflict, although the coverage from surrounding countries remains good ensuring some level of ground constraint on the satellite measurements.

Precipitation anomalies are presented relative to the 2004-2009 period, for consistency with the GRACE anomalies. To aid interpretation of the time-series graphs it is also necessary to provide the location of the 2001-2020 average anomaly, i.e. where the ‘zero’ line would sit if we were to reference rainfall to this longer and more representative timeframe: Western Yemen (9 mm), Ramlat (5 mm), eastern highlands (7 mm), Abyan (9 mm), central highland plains (17 mm), north west (6 mm), southern Tihamah (13 mm), southern deltas (12 mm), northern highland plains (10 mm), Hajjah (23 mm), western mountains (12 mm).

The response of groundwater storage to precipitation is a complex and active science. Aquifers comprise a temporally sensitive short response and insensitive long response. By the nature of the short GRACE time-series we can only look at the short responses here, but as groundwater use in Yemen is predominantly from shallow and rechargeable basins, this is possible. Indeed, water table ratio values are above 1 in the much of western Yemen, indicating aquifers with a coupling to land processes and climate dynamics. This area is relatively humid, and in similar conditions Opie et al., (2020), report strong monthly correlations between groundwater storage and precipitation. Similar relationships have been identified in the US. Hence, we followed a similar approach to Opie et al., and found a reasonable correlation between monthly groundwater and storage for all of western Yemen and some individual regions. We use these relationships to approximate the expected groundwater anomaly based on precipitation changes. We recognise this is a very crude analysis and subject to significant uncertainty. Nonetheless, it helps highlight further the significance of the discrepancy between high precipitation totals and lowering groundwater in the second half of the conflict.

F. Conflict data

Conflict incidents data from the Armed Conflict Location & Event Data Project (ACLED). In the key metrics time-series plots incidents are also presented as a percentage of all incidents to minimise differences in reporting coverage during the conflict.

The notes for each entry into the ACLED database for Yemen between Jan 2015 and Dec 2020 were filtered using keywords to find events associated with the energy, agricultural and water sectors.

To avoid false positives, we also searched for exemptions, e.g. we identify entries with ‘field’ but then blacklist these if the field is ‘oil field’, ‘field commander’, ‘field marshal’. The filtered entries have undergone some, but not exhaustive manual verification.

For the agriculture filtering, there is some complication surrounding the Yemeni Dairy Factory. This is a location in Hodeidah close to lots of fighting, and as such, many events are coded as occurring there, even if the dairy factory was not the direct target or location of fighting. The incidents explicitly labelled as coded at the dairy factory have been filtered out, but 21 incidents in which the dairy factory is mentioned but not explicitly as the coded location remain included. For example, incidents like “Anti-Houthi forces including from the Saudi-led coalition reportedly fired heavy and medium weapons at residential and commercial areas in Sana’a Street next to the Yamani Dairy factory in Hodeidah city western Yemen with no report of casualties” remains in the filtered results.

The keyword list for water is:

‘water’; ‘wells’; ‘pump’; ‘sewage’; ‘reservoir’; ‘drink’; ‘irrigation’; ‘wastewater’; ‘desalination’; ‘irrigate’; ‘drinking’; ‘weir’; ‘spate’; ‘sanitation’; ‘dam’; ‘LWSC’; ‘GARWSP’; ‘MWE’; ‘NWRA’; ‘WASH’; ‘treatment plant’; ‘flood’.

And the exemption list:

‘oil’; ‘Al Wash’; ‘blackwater’.

The keyword list for energy is:

‘electric’; ‘diesel’; ‘solar’; ‘oil well’; ‘oil pump’; ‘substation’; ‘sub-station’; ‘energy grid’; ‘grid’; ‘gas’; ‘transformer’; ‘fuel’; ‘filling station’; ‘power plants’; ‘power lines’; ‘energy’; ‘electrical’; ‘lights’; ‘generator’; ‘transmission line’; ‘utilities’; ‘crude’; ‘pipeline’; ‘battery’; ‘pylon’; ‘coal’; ‘firewood’; ‘power cut’; ‘power outage’; ‘fuel tanks’; ‘safer’; ‘oil company’; ‘oil facility’; ‘oil pipe’; ‘oil terminal’; ‘petrol’; ‘lng’; ‘lpg’; ‘refinery’; ‘ypg’.

And the exemption list:

‘tear gas’; ‘water pipeline’; ‘fuelling’.

The keyword list for agriculture is:

‘soil’; ‘crop’; ‘agriculture’; ‘farm’; ‘fish’; ‘fisherman’; ‘fisheries’; ‘bee’; ‘beekeeper’; ‘plants’; ‘salt’ ; ‘field’; ‘agri’; ‘agro’; ‘vegetable’; ‘flower’; ‘plantation’; ‘onion’; ‘tomato’; ‘okra’; ‘legume’; ‘zucchini’; ‘mulukhiyah’; ‘tree’; ‘qat’; ‘fruit’; ‘coffee’; ‘palm’; ‘banana’; ‘mango’; ‘pomegranate’; ‘melon’; ‘cotton’; ‘orchard’; ‘garden’; ‘arbor’; ‘tractor’; ‘polytunnel’; ‘nursery’; ‘greenhouse’; ‘dairy’; ‘milk’; ‘livestock’; ‘graze’; ‘grazing’; ‘sheep’; ‘goat’; ‘chicken’; ‘cattle’; ‘cow’; ‘cows’; ‘donkey’; ‘donkeys’; ‘poultry’; ‘camel’; ‘horse’; ‘ruminant’; ‘forage’; ‘foraging’; ‘pasture’; ‘herd’; ‘sorghum’; ‘millet’; ‘corn’; ‘cereal’; ‘barley’; ‘maize’; ‘sesame’; ‘Tihama Development Authority’; ‘fertilizer’; ‘pesticide’; ‘tda’.

And the exemption list:

‘seller’; ‘market’; ‘dealer’; ‘harbour’; ‘oil field’; ‘field leader’; ‘field commander’; ‘Raydan field’; ‘hospital’; ‘coded at the Yamani Dairy Factory’; ‘coded at the Yemeni Dairy Factory’; ‘coded at the Yaman Dairy Factory’; ‘coded at the Yemen dairy factory’; ‘coded at the Al Yamani Dairy Factory’; ‘coded at the Al Yemeni Dairy Factory’; ‘coded at the Al Yaman Dairy Factory’; ‘coded at the Al Yemen dairy factory’; ‘Yamani Dairy Factory]’; ‘Yemeni Dairy Factory]’; ‘Yaman Dairy Factory]’; ‘Yemen dairy factory]’; ‘Al Yamani Dairy Factory]’; ‘Al Yemeni Dairy Factory]’; ‘Al Yaman Dairy Factory]’; ‘Al Yemen dairy factory]’; ‘coded at Yamani Dairy Factory’; ‘coded at Yemeni Dairy Factory’; ‘coded at Yaman Dairy Factory’; ‘coded at Yemen dairy factory’; ‘coded at Al Yamani Dairy Factory’; ‘coded at Al Yemeni Dairy Factory’; ‘coded at Al Yaman Dairy Factory’; ‘coded at Al Yemen dairy factory’; ‘coded at the Yamani Dairy Factory’; ‘coded at the Yemeni Dairy Factory’; ‘coded at the Yaman Dairy Factory’; ‘coded at the Yemen dairy factory’; ‘coded at the Al Yamani Dairy Factory’; ‘coded at the Al Yemeni Dairy Factory’; ‘coded at the Al Yaman Dairy Factory’; ‘coded at the Al Yemen dairy factory’; ‘coded in Yamani Dairy Factory’; ‘coded in Yemeni Dairy Factory’; ‘coded in Yaman Dairy Factory’; ‘coded in Yemen dairy factory’; ‘coded in Al Yamani Dairy Factory’; ‘coded in Al Yemeni Dairy Factory’; ‘coded in Al Yaman Dairy Factory’; ‘coded in Al Yemen dairy factory’; ‘coded in the Yamani Dairy Factory’; ‘coded in the Yemeni Dairy Factory’; ‘coded in the Yaman Dairy Factory’; ‘coded in the Yemen dairy factory’; ‘coded in the Al Yamani Dairy Factory’; ‘coded in the Al Yemeni Dairy Factory’; ‘coded in the Al Yaman Dairy Factory’; ‘coded in the Al Yemen dairy factory’; ‘from Yamani Dairy Factory’; ‘from Yemeni Dairy Factory’; ‘from Yaman Dairy Factory’; ‘from Yemen dairy factory’; ‘from Al Yamani Dairy Factory’; ‘from Al Yemeni Dairy Factory’; ‘from Al Yaman Dairy Factory’; ‘from Al Yemen dairy factory’; ‘city max’;

Return to Contents