Africa RiskView END OF SEASON REPORT | (2016)

This Africa RiskView End of Season Report is a publication by the African Risk Capacity (ARC). The report discusses Africa RiskView’s estimates of rainfall, drought and population affected, comparing them to information from the ground and from external sources. It also provides the basis of a validation exercise of Africa RiskView, which is conducted in each country at the end of an insured season. This exercise aims at reviewing the performance of the model and ensuring that the country’s drought risk is accurately reproduced by Africa RiskView for drought monitoring and insurance coverage.

Highlights:

Rainfall: Affected Populations:  Overall good performance of the 2016 rainy season, with nor-  Africa RiskView estimates that only some localised areas in mal to above normal cumulative rainfall totals at national and central and south-western Niger were affected by drought regional. However, localised areas in central Niger experienced conditions. Overall, around 570,000 people were estimated to below normal and erratic rainfall from August to October. be affected in and Dosso regions, which remains well Drought: below the modelled historical average of around 1.3 million  Despite an overall good performance of the 2016 agricultural people. season according to Africa RiskView, some areas in central and ARC Risk Pool: south-western Niger experienced localised dry conditions  Due to the overall good performance of the season and de- which resulted in below normal WRSI values. The end-of- spite localised drought impacts in central and south-western season WRSI is higher than normal in western and eastern Niger, the trigger for a payout from ARC Ltd was not reached Niger. at the end of the 2016 agricultural season.

Rainfall above normal rainfall in May and June, particularly in south- The rainy season in Niger lasts from May to October. The 2016 western Niger. This was followed by a normal progression of the season was characterised by the expected high variability of rains throughout the country. However, the rains were slightly cumulative rainfall received across the different regions of the below normal in region in August, and in south-eastern country. In the south-western parts of the country, namely in Niger ( and Zinder regions) in late July. It is important to note Dosso and Tillabéri regions, as well as in the southern parts of that at sub-regional level, the performance of the seasonal rains in central Niger, up to over 600 mm of rainfall were was more mixed. For instance, in Tarka and départements received over the course of the 2016 season. Conversely, agricul- in northern , below normal rainfall was received in tural areas further north (Tahoua and Zinder regions) and east late August and September, with an early cessation of the season. () only received between 300 and 600 mm of rains. Similarly, in Daragaram Takaya, , and Takiéta Compared to the long-term average (1983-2015), the cumulative départements in Zinder region experienced erratic and below seasonal rains were normal to above normal throughout the normal rainfall, particularly from July onwards. country. In the northern parts of Tillabéri region in south-western Drought Niger, over 150% of average rainfall was received; the same ap- The in-country Technical Working Group (TWG) customised Africa plies to most of Diffa region in the country’s south-east. However, RiskView to model the impact of drought on , one of the localised areas in central Zinder region received slightly below main staple crops in Niger. Based on the parameters chosen normal rainfall (around 80% of average). during the customisation of the model, Africa RiskView models all Regarding the spatial and temporal distribution of the seasonal possible planting opportunities during the planting window (May- rains in 2016, an analysis of dekadal (10-day) rainfall at the region- July), and chooses the one which results in the best WRSI value at al level suggests that the season started early with normal to the end of the season. In 2016, Africa RiskView estimates that in

For more information visit our website: www.africanriskcapacity.org Africa RiskView END OF SEASON REPORT | NIGER (2016)

Cumulative rainfall in mm, Cumulative rainfall in % of 1983-2015 average, May-Oct 2016, Niger (ARC2) May-Oct 2016, Niger (ARC2)

End-of-season WRSI, Niger, End-of-season WRSI compared to 10-year average, Niger, 2016 agricultural season 2016 agricultural season most areas, the best planting conditions were reached in late June According to this information, sowing started in May 2016 in the (21-30 June). In some parts of southern and central Niger, planting south-western parts of Niger, with over one third of villages in occurred slightly earlier than this (11-20 June), while in other Dosso, Tillabéri and Tahoua recording successful planting, particu- areas, particularly northern Dosso and southern Tahoua, planting larly during the last dekad of the month (21-31 May). In Maradi occurred only in early July (1-10 July), according to Africa and Zinder regions in central Niger, planting occurred mostly in RiskView. Compared to normal planting conditions (based on the the last two dekads of June (11-30 June), while in Diffa, one third 1983-2015 average), normal planting occurred in most of western of villages planted in late June, while the rest planted in early to Niger (Tillabéri and Tahoua regions), but slightly later than normal mid-July. It is important to keep in mind that there is a fundamen- (up to 1 dekad) in northern Dosso and southern Tahoua. Through- tal methodological difference in observed planting dates and the out central and eastern Niger, Africa RiskView suggests that planting dates modelled by Africa RiskView, which calculates the planting started 1-2 dekads earlier than normal. best planting dekad based on the rainfall threshold set by the TWG (15 mm of rain in one dekad). Data on planting dates collected in the field by the Ministry of Agriculture also indicates that planting started earlier than normal The end-of-season WRSI values highlight a high degree of variabil- in 2016, due to the favourable rainfall received in May and June. ity across the country, which is in line with normal rainfall patterns

For more information visit our website: www.africanriskcapacity.org Africa RiskView END OF SEASON REPORT | NIGER (2016)

Estimated population affected by drought, Estimated population affected by drought, Niger, 2016 agricultural season Niger, 1983-2016 in the country. In south-western Niger, and in the southern parts Affected Populations of Maradi region in central Niger, the WRSI values tend to be Based on the in-country customisation of Africa RiskView, around higher, and the water requirements of the reference crop (millet) 6.1 million people in Niger are vulnerable to drought. At the end were fully satisfied at the end of the 2016 agricultural season. The of the 2016 agricultural season, Africa RiskView estimates that WRSI values gradually decrease towards northern and eastern around 570,000 people were directly affected by drought condi- Niger, where agricultural production and yields tend to be lower tions in central and south-western Niger. The vast majority of than in the more fertile south-west. Based on the FEWS NET WRSI drought affected people were in Zinder region, namely in Tanout classification, mediocre WRSI values (60-80) prevailed in these (204,000 people), Mirriah (125,000), Takiéta (89,000), Magaria areas in 2016. (77,000) and (29,000) départements. Moreo- Compared to the benchmark set by the TWG to model normal ver, an estimated 34,000 people in and 13,000 in Dosso conditions in Niger (average of the previous 10 years), it appears départements of were estimated to be affected by that the end-of-season WRSI values are above normal in usually drought conditions at the end of the 2016 season. Compared to drier areas such as northern Tahoua, eastern Zinder, and Diffa historical drought years modelled by Africa RiskView, it appears regions, as well as in north-western Tillabéri region. However, that the total number of people affected by drought in Niger in localised areas in Dosso and western Zinder experienced below 2016 was well below the historical average of around 1.3 million normal WRSI conditions. These were mostly driven by a higher people. The major droughts in Niger occurred in the 1980s, mid- than normal water deficit during the flowering phase in Zinder, 1990s, as well as in 2004 and 2009, as well as more recently in and during the ripening phase of the reference crop in Dosso. 2014 and to a lesser extent in 2011, according to Africa RiskView.

Africa RiskView’s WRSI calculations are similar to FEWS NET’s The 2016 Cadre Harmonisé exercise, concluded in November regional WRSI model, which indicates that mediocre to average 2016, found that nearly 330,000 people were in severe food WRSI values were prevailing in Niger at the end of October 2016, insecurity (Phase 3 and above) at the time of the analysis. This with a slightly better outlook for the southern parts of Tahoua figure was estimated to rise to nearly 750,000 people during the region in central Niger. However, FEWS NET’s WRSI anomaly map peak lean season (June-August 2017). According to the analysis, projects a less optimistic view of the situation in south-western three départements, namely Bani Bangou in northern Tillabéri, and central Niger, with large parts of Tillabéri, Dosso and Zinder in eastern Maradi and N’Gourti in Diffa region could fall having below normal WRSI values. FAO’s Agricultural Stress Index into Crisis due to the effect of a depletion of stocks, poor terms of (ASI), a composite indicator based on vegetation and temperature trade between livestock and cereals and the impact of civil insecu- information, suggests that the situation for all cropped areas in rity on markets and humanitarian access. In addition, food insecu- Niger was positive throughout the 2016 season. rity in Diffa region is expected to be driven by the impact of the

For more information visit our website: www.africanriskcapacity.org Africa RiskView END OF SEASON REPORT | NIGER (2016)

Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 -2015 average (blue line),May-Oct 16, -2015 average (blue line), May-Oct 16, -2015 average (blue line), May-Oct 16, Tillabéri, Niger Dosso, Niger Tahoua, Niger

Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 -2015 average (blue line), May-Oct 16, -2015 average (blue line), May-Oct 16, -2015 average (blue line), May-Oct 16, Maradi, Niger Zinder, Niger Diffa, Niger

conflict in the Lake region. It appears thus that apart from RiskView in view of Niger’s participation in the 2017/18 ARC Risk some localised areas, drought was not a major driver of food Pool. The exercise aims at reviewing the drought index parame- insecurity in Niger at the end of the 2016 agricultural season. ters used by the model, as well as updating input data such as the vulnerability profile and poverty information used by Africa ARC Risk Pool RiskView. Potential improvements to the model will help ensure Niger has been a member of the ARC Risk Pool since the first pool that drought risks are accurately reproduced for drought monitor- in 2014/15. In that year, the country received a payout of around ing and insurance coverage and that the modelling continues to USD 3.5 million, due to the poor performance of the 2014 agricul- evolve as new information is reported and gathered. tural season in West Africa. During the current pool, the country did not receive for a payout from their policy as the attachment level selected by the Government of Niger (around 2.8 million people) was not reached.

The in-country Technical Working Group with support from the ARC Secretariat is currently reviewing the customisation of Africa

For more information visit our website: www.africanriskcapacity.org Africa RiskView END OF SEASON REPORT | NIGER (2016)

Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 -2015 average (blue line), May-Oct 16, -2015 average (blue line), May-Oct 16, -2015 average (blue line), May-Oct 16, Tarka, Zinder, Niger Tanout, Zinder, Niger Takiéta, Zinder, Niger

Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 Dekadal rainfall in mm compared to 1983 -2015 average (blue line), May-Oct 16, -2015 average (blue line), May-Oct 16, -2015 average (blue line), May-Oct 16, D. Takaya, Zinder, Niger Mirriah, Zinder, Niger Magaria, Zinder, Niger

For more information visit our website: www.africanriskcapacity.org Africa RiskView END OF SEASON REPORT | NIGER (2016)

About ARC: The African Risk Capacity (ARC) is a special- The Africa RiskView software is the tech- The ARC Insurance Company Limited is the ised agency of the African Union designed nical engine of ARC. It uses satellite-based financial affiliate of the ARC Agency, which to improve the capacity of AU Member rainfall information to estimate the costs of pools risk across the continent through issu- States to manage natural disaster risk, responding to a drought, which triggers a ing insurance policies to participating coun- adapt to climate change and protect food corresponding insurance payout. tries. insecure populations.

Note on Africa RiskView’s Methodology:

Rainfall: Africa RiskView uses Drought: Africa RiskView uses Affected Populations: Based on Response Costs: In a fourth various satellite rainfall da- the Water Requirements Satis- the WRSI calculations, Africa and final step, Africa RiskView tasets to track the progression faction Index (WRSI) as an indi- RiskView estimates the number converts the numbers of affect- of rainy seasons in Africa. Coun- cator for drought. The WRSI is of people potentially affected ed people into response costs. tries intending to participate in an index developed by the Food by drought for each country For countries participating in the ARC Risk Pool are required and Agriculture Organisation of participating in the insurance the insurance pool these na- to customise the rainfall com- the United Nations (FAO), pool. As part of the in-country tional response costs are the ponent by selecting the dataset which, based on satellite rain- customisation process, vulnera- underlying basis of the insur- which corresponds the best to fall estimates, calculates wheth- bility profiles are developed at ance policies. Payouts will be the actual rainfall measured on er a particular crop is getting the sub-national level for each triggered from the ARC Insur- the ground. the amount of water it needs at country, which define the po- ance Company Limited to coun- different stages of its develop- tential impact of a drought on tries where the estimated re- ment. To maximise the accura- the population living in a spe- sponse cost at the end of the cy of Africa RiskView, countries cific area. season exceeds a pre-defined intending to take out insurance threshold specified in the insur- customise the software’s pa- ance contracts. rameters to reflect the realities on the ground.

Disclaimer: The data and information contained in this report have been developed for the purposes of, and using the methodology of, Afri ca RiskView and the African Risk Capacity Group. The data in this report is provided to the public for information purposes only, and neither the ARC Agen- cy, its affiliates nor each of their respective officers, directors, employees and agents make any representation or warranty regarding the fitness of the data and information for any particular purpose. In no event shall the ARC Agency, its affiliates nor each of their respective officers, directors, employees and agents be held liable with respect to any subject matter presented here. Payouts under insurance policies issued by ARC Insurance Company Limited are calculated using a stand-alone version of Africa RiskView, the results of which can differ from those presented here.

For more information visit our website: www.africanriskcapacity.org