Climate Profile May27 - Makerere Universitymuccri.mak.ac.ug/sites/default/files/Publications/Climate... · 2016-04-07 · iii! (ACKNOWLEDGMENTS( This!report!was!prepared!by!Megan!Lickley!(Climate!Specialist,!World!Bank!consultant),!supported!by!Halla!

!

!

!

Uganda&Climate(Profile((

World(Bank,(May(1,(2015(

!((((((((((((((((((((((((((

!

!

i!ABBREVIATIONS(AND(ACRONYMS((/& & Per(%& & Percentage(°" " Degrees!°C" " Degrees!Centigrade!AR5" " Fifth!Assessment!Report!(of!the!IPCC)!BCSD" " Bias:Corrected!and!Spatially!Disaggregated!CMI" " Climate!Moisture!Index!CMIP3"(or"5)" Coupled!Model!Intercomparison!Project!Phase!Three!(or!Five)!CV" " Coefficient!of!Variation!DfID" " Department!of!International!Development!(UK)!GCM" " Global!Circulation!Model!HWDI" " Heat!Wave!Duration!Index!IPCC" " Intergovernmental!Panel!on!Climate!Change!JJA" " June,!July,!August!MAM" " March,!April,!May!mm" " Millimeter!NDP"II" " Second!National!Development!Plan!PET" " Potential!Evapotranspiration!RCP" " Representative!Concentration!Pathway!SON" " September,!October,!November!SRES" " Special!Report!on!Emissions!Scenarios!UNFCC"" United!Nations!Framework!Convention!USAID" " United!States!Agency!for!International!Development!W/m2! ! Watts!per!meter!squared!WASP" " Weighted!Anomaly!Standardized!Precipitation!Index!!

!

!

ii!TABLE(OF(CONTENTS(((Acknowledgements(……….………………….………………….………………….………………….………………………..( iii(Executive(Summary(……….………………….………………….………………….………………….………………………..( iv(Introduction(…….…………….………………….………………….………………….………………….………………………..( 1(Previous(Work(……………….………………….………………….………………….………………….………………………..( 3(Methodology((…………………………………………………………………………………………………………………………( 3(Historical(Climate(and(Trends(…………….………………….………………….………………….………………..........( 6(Climate(Change(Projection(Overview(…..………………….………………….………………….………………………( 8(Climate(Projection(By(Indicator(…………..………………….………………….………………….………………..(…….( 10(( Mean(Temperature(……………………..……………….………………….………………….……………………..( 11(( Maximum(Daily(Temperature(.………………….………………….………………….………………………….( 12(( Minimum(Daily(Temperature(.………………….………………….………………….…………………………..( 13(( Heat(Wave(Duration(Index(…..………………….………………….………………….……………………(……..( 14(( Simple(Daily(Precipitation(Index((.………………….………………….………………….………………(……..( 15(( 5WDay(Precipitation(…………...………………….………………….…………………….…………………………..( 16(( Consecutive(Number(of(Dry(Days(.………………….………………….…………………………………(……..( 17(( Monthly(Precipitation(……..………………….………………….………………….……………………………….( 18(( Frequency(of(Precipitation(Extremes(.………………….………………….……….………………………….( 19(( Potential(Evapotranspiration(.………………….………………….………………….…………………………..( 20(( Weighted(Anomaly(Standardized(Precipitation(Index((…….……………….………………………….( 21(( DecWFeb(Monthly(Precipitation(.………………….………………….………………….………………………..( 22(( MarWMay(Monthly(Precipitation(.………………….………………….………………….……..……………….( 23(( JunWAug(Monthly(Precipitation.………………….………………….………………….…………………………( 24(( SepWNov(Monthly(Precipitation(………………….………………….………………….…………….………….( 25(( Number(of(Wet(Days(A(Year(.………………….………………….………………….………………..………….( 26(( Climate(Moisture(Index(.………………….………………….………………….…………………………….…….( 27(Conclusions(.………………….………………….………………….………………………………………………………………..( 28(References(……………………………………………………………………………………………………………………………..( 30(Annex(A(………………………………………………………………………………………………………………………………….( 31(Annex(B(………………………………………………………………………………………………………………………............( 32(Annex(C(………………………………………………………………………………………………………………………………….( 33(Annex(D(………………………………………………………………………………………………………………………………….( 36(Annex(E(…………………………………………………………………..……………………………………………………………..((((((40(

(((!

!

!

iii!(ACKNOWLEDGMENTS((This!report!was!prepared!by!Megan!Lickley!(Climate!Specialist,!World!Bank!consultant),!supported!by!Halla!Qaddumi!(Natural!Resources!Economist,!World!Bank!consultant),!Brent!Boehlert!(Senior!Associate,!Industrial!Economics),!and!Tony!Garvey!(Strategic!Water!Resources!Advisor,!World!Bank!consultant),!under!the!leadership!of!Berina!Uwimbabazi!(Sr.!Water!Resources!Management!Specialist!and!TTL,!GWADR).!!The team would like to!thank!the!reviewers!who!provided!valuable!comments!and!suggestions!that!improved!the!quality!of!the!report,!including!Ana!Elisa!Bucher!(Climate!Change!Specialist,!GCCPT)!and!Christoph!Pusch!(Lead!Disaster!Risk!Management!Specialist,!GSURR!on!behalf!of!Africa!Disaster!Risk!Management).!!The!report!was!prepared!in!collaboration!with the Government of Uganda, Climate Change Department in Ministry of Water and Environment and Makerere University.!!DFID!funding!is!gratefully!acknowledged.!

!

!

iv!EXECUTIVE(SUMMARY((

INTRODUCTION((

This!report!analyzes!and!presents!historical!and!future!climate!trends!in!Uganda!using!a!collection!of!the!most!recent!climate!projection!information!available.!!This!report!represents!the!most!in:depth!climate!analysis!to!date!for!Uganda,!building!on!a!number!of!previous!efforts.!!The!aim!is!to!provide!a!rigorously!derived!and!comprehensive!set!of!information!on!historical!and!future!climate!to!inform!climate:resilient!development!planning!for!the!country.!!!(

METHODOLOGY((

This!analysis!uses!seventeen!key!indicators!to!characterize!historical!and!future!climate!and!results!are!presented!for!each!indicator!in!order!to!provide!a!comprehensive!picture.!!Trends!of!historical!climate!are!based!on!analysis!of!a!global!public!domain!dataset!from!the!Princeton!Land!Surface!Hydrology!Research!Group.!!Trends!of!projected!future!climate!are!based!on!analysis!of!a!subset!of!the!multi:model!ensemble!from!the!most!recent!Intergovernmental!Panel!on!Climate!Change!(IPCC)!Assessment!Report,!known!as!the!Fifth!Assessment!Report!(AR5)!(see!Table"3!for!a!summary!of!the!models!used).!!Two!Representative!Concentration!Pathways!(RCPs)!are!considered!in!the!analysis,!one!moderate!emissions!scenario!(RCP!4.5),!where!temperatures!stabilize!in!the!second!half!of!the!21st!century,!and!one!high!emissions!scenario!(RCP!8.5),!where!temperatures!continue!to!increase!throughout!the!21st!century.!!Models!are!bias!corrected!to!be!statistically!consistent!with!the!historical!data!set!and!include!temperature!and!precipitation!output,!down:scaled!to!a!0.5°!x!0.5°"resolution.!!Results!of!the!analysis!focus!on!two!time!periods!for!future!projections:!mid:century!(2030:2050)!and!end:of:century!(2080:2100).!!!Current!and!future!climate!varies!not!only!over!time,!but!also!over!space,!with!some!regions!projected!to!see!greater!changes!in!precipitation!or!temperature!than!others.!!Thus,!climatic!conditions!are!presented!in!this!report!on!both!a!national!scale!and!a!sub:national!scale,!at!the!level!of!the!Water!Management!Zone!(WMZ).!!!(

KEY(FINDINGS((

Historical*Climate*Historically,!Uganda!has!had!a!highly!variable!climate!with!average!annual!temperature!ranging!from!below!16°C!in!some!regions!to!as!high!as!25°C!in!others.!!Average!rainfall!ranges!between!500!mm/year!and!2100!mm/year!across!the!country,!with!two!rainy!seasons!from!March:May!and!from!September:November.!!There!is!a!high!level!of!interannual!variability!in!monthly!precipitation.!!The!most!extreme!monthly!precipitation!events!occurred!in!1962!and!1997,!two!years!when!extensive!flooding!was!witnessed!in!various!parts!of!Uganda.!!

!

Monthly(mean(temperature(from(1950W2008( Monthly(mean(precipitation(from(1950W2008(

HISTORICAL&

Box(1:(Historical(Temperature(and(precipitation(

!

!

v!Trends*in*Climate!!Temperatures!have!been!increasing!in!Uganda!by!approximately!0.2°C/decade!over!the!last!30!years.!!!There!have!been!no!statistically!significant!changes!in!annual!rainfall!to!date.!!However,!recent!studies!have!found!a!decrease!in!rainfall!during!the!March:May!rainy!season.!!!Box!1!shows!historical!precipitation!and!temperature!over!the!period!1950!to!2008.!!*Temperature*Projections*Models!project!warming!to!continue!to!increase!throughout!the!21st!century!and!to!be!between!2!and!6°C!warmer!!than!historical!temperature!by!the!end!of!the!century.!!Projections!show!that!there!will!be!an!increase!in!minimum!daily!temperature!and!an!even!greater!increase!in!maximum!daily!temperature.!The!

analysis!also!indicates!an!increase!in!both!frequency!and!duration!of!heat!waves.!!!As!a!result!of!warmer!temperatures,!higher!levels!of!evaporation!are!projected,!leading!to!an!increase!in!aridity!in!the!country.!!!A!summary!of!temperature!projections!and!extremes!is!provided!in!Box!2.!*Precipitation*Projections*There!is!less!agreement!amongst!models!on!the!direction!and!magnitude!of!the!change!in!future!average!annual!precipitation.!!However,!models!agree!that!precipitation!extremes!will!increase!and!that!there!will!be!an!increase!in!variability.!!!!The!majority!of!the!models!project!that!the!highest!annual!5:day!precipitation!levels!will!increase!by!at!least!50%.!!By!mid:century,!monthly!precipitation!extremes!are!projected!to!exceed!historical!extremes!both!in!frequency!and!magnitude.!!Most!models!project!an!increase!in!rainfall!during!the!December:February!dry!season!and!a!decrease!in!rainfall!during!the!March:May!rainy!season.!!!While!precipitation!extremes!are!expected!to!increase,!so!too!is!the!duration!of!consecutive!dry!days;!models!project!the!longest!annual!duration!of!dry!days!to!increase!by!approximately!one!week!by!mid:century.!!A!summary!of!precipitation!projections!is!provided!in!Box!3.!

Box(2:(Summary(plots(of(temperature(projections(

TEMPERATURE&

TEMPERATURE&EXTREMES&

MultiWmodel(temperature(projections(

!0°C! ! ! !!!!!!!!!!!6°C!

RCP(4.5( ( ((RCP(8.5(

2080

W210

0((

2030

W205

0(

Median(model(temperature(change(

The(median(model(projection(of(the(number(of(days(annually(where(maximum(temperature(exceeds(

RCP(4.5(

2030

W205

0(

Historical(

RCP(8.5(

2080

W210

0(

RCP(8.5(

2030

W205

0(

RCP(4.5(

2080

W210

0(

300(((

200(((

100((

(

0(

!

!

vi!

(

CONCLUSIONS((

Looking!across!the!findings,!some!broad!conclusions!can!be!drawn.!!Temperatures!have!already!begun!to!increase!and!will!continue!to!increase!throughout!the!century.!!Uganda!has!historically!had!high!variability!in!precipitation,!which!is!projected!to!increase!in!the!future.!!!Projections!show!that!there!will!be!more!extreme!events,!in!terms!of!both!more!frequent!acute!rainfall!events!and!more!frequent!and!longer!dry!periods.!!With!increasing!temperatures!and!increasing!variability!in!rainfall,!Uganda!is!expected!to!become!more!arid!in!the!northern!and!eastern!parts!of!the!country.!!The!Lake!Victoria!region,!however,!is!expected!to!become!wetter.!!!

This!report!builds!on!existing!work!by!using!a!number!of!indicators!and!the!most!recent!information!available!to!more!fully!characterize!historical!and!future!climatic!conditions!in!Uganda.!!It!provides!insights!into!expected!changes!that!are!likely!to!occur!in!the!future.!!This!climate!profile!is!not!a!climate!impact!or!risk!assessment;!more!work!would!be!needed!to!translate!the!results!of!this!analysis!into!impacts!and!risks.!!Notwithstanding!the!rigor!of!the!analysis,!there!is!also!substantial!uncertainty!about!the!state!of!future!climate.!!The!results,!however,!clearly!show!that!Uganda!is!already!contending!with!a!high!degree!of!climate!variability,!across!both!space!and!time,!and!that!this!is!expected!to!increase!in!the!future.!!The!implication!is!that!development!planning!must!carefully!consider!and!fully!integrate!the!variability!of!today’s!climate,!as!it!presents!challenges!that!have!yet!to!be!adequately!addressed.!!Adapting!to!current!climatic!conditions!will!prepare!Uganda!to!be!able!to!manage!potentially!more!complex!–!and!uncertain!–!climate:related!risks!in!the!future.!

Box(3:(Summary(plots(of(precipitation(projections(

PRECIPITATION&

PRECIPITATION&EXTREMES&

MultiWmodel(precipitation(projections(

2080

W210

0((

2030

W205

0(

RCP(4.5( ( ((RCP(8.5(

(0( (((((((((((20( ( ((40(

Median(model(precipitation(change((grey(means(<66%(of(models(agree(on(direction(of(change)(

RCP(4.5(

2030

W205

0(

Historical(

RCP(8.5(

2030

W205

0(

RCP(8.5(

2080

W210

0(

RCP(4.5(

2080

W210

0(W(300((

((W(600(

The(50Wyear(monthly(rainfall(event(in(mm/month(

!

!

1!INTRODUCTION!!This!report!focuses!on!historical!climate!and!future!climate!in!Uganda,!both!nationally!and!at!the!level!of!the!Water!Management!Zone!(WMZ).!!The!aim!is!to!analyze!and!present!the!most!complete!and!recent!information!available!on!historical!and!future!climate!trends!to!support!climate:resilient!development!planning!for!the!country.!!Current!climatic!conditions!are!a!critical!component!of!development,!and!climate!change!adds!to!the!variability!and!unpredictability,!presenting!new!ranges!of!risk!(see!Box!1).!!!!!This!report!presents!analysis!of!both!historical!climate!and!future!climate!projections!with!reference!to!several!key!indicators!derived!from!daily!minimum,!maximum!and!mean!temperature,!and!monthly!and!daily!precipitation.!!These!indicators!are!summarized!in!Table"1!and!are!used!to!characterize!and!compare!climate!scenarios!relative!to!historical!climate.!!!!"Table"1:"Indicators!used!to!characterize!historic!and!projected!climate!

Indicator! Description!Mean"Daily"Temperature" Temperature!temporally!averaged!for!each!day.!(°C)!Minimum"Daily"Temperature" Minimum!temperature!for!each!day.!(°C)!Maximum"Daily"Temperature" Maximum!temperature!for!each!day.!(°C)!Heat"Wave"Duration"Index"(HWDI)"

The!number!of!consecutive!days!where!daily!temperature!exceeds!the!climatological!daily!temperature!by!5°C!or!more.!!!

Potential&Evapotranspiration&(PET)" The!evaporation!that!would!occur!if!a!sufficient!source!of!water!were!available.!(mm)!Climate"Moisture"Index"(CMI)" An!indication!of!aridity!derived!from!annual!precipitation!and!PET!and!ranging!from!!

:1!to!1,!where!negative!values!indicate!dryness!and!positive!values!indicate!wetness.!!!!

Mean"Monthly"Precipitation" Total!rainfall!averaged!by!month.!(mm/month)!DJF&monthly&precipitation& Monthly!rainfall!during!December,!January,!and!February!(DJF).!(mm)!&MAM&monthly&precipitation& Monthly!rainfall!during!the!March,!April,!May!(MAM)!season.!(mm)!JJA&monthly&precipitation& Monthly!rainfall!during!June,!July,!August!(JJA).!(mm)!SON&monthly&precipitation& Monthly!rainfall!during!September,!October,!November.!(mm)!Number&of&wet&days&per&year& The!number!of!days!a!year!where!precipitation!exceeds!1mm.!Frequency&monthly&precip&>200&mm& Annual!frequency!where!monthly!precipitation!exceeds!200mm/month.!!!Frequency&monthly&precip&>250&mm& Annual!frequency!where!monthly!precipitation!exceeds!250mm/month.!!!!Frequency&monthly&precip&>300&mm& Annual!frequency!where!monthly!precipitation!exceeds!300mm/month.!!!Max&number&of&consecutive&dry&days& Number!of!consecutive!days!a!year!where!precipitation!is!less!than!1mm.!!Maximum&precip&over&5Tdays&& Maximum!rainfall!over!a!5:day!period!each!year!(indication!of!extreme!rainfall).!

(mm)!Weighted&Anomaly&Standardized&Precipitation&Index&(WASP)&

Indication!of!how!monthly!rainfall!deviates!from!historical!rainfall.!!Positive!values!indicate!above!normal!rainfall,!negative!values!indicate!below!normal!rainfall.!!!!

Simple&daily&precipitation&Index&& Average!rainfall!during!a!wet!day,!where!a!wet!day!occurs!when!precip!!>!1mm.!!!

Box(1:((Climate(Change(versus(Variability((

Climate(varies(naturally(on(all(timeWscales(–(season(to(season,(year(to(year,(decade(to(decade(and(even(on(longerWtime(scales(–(because(the(components(are(never(in(equilibrium(and(are(constantly(shifting.((These(fluctuations(do(not(represent(climate(change.(Climate(change(is(rather(a(shift(in(the(statistical(properties(of(the(climate(system(when(considered(over(long(periods(of(time.(It(is(identified(by(changes(in(the(mean(and(/(or(variability(of(climate(properties(that(persist(for(an(extended(period,(typically(decades(or(longer.((

There(is(consensus(among(the(scientific(community(that(there(has(been(an(unequivocal(warming(of(the(climate(system(since(1950.((Further,(it(is(extremely(likely(that(most(of(the(observed(warming(over(this(time(period(has(been(caused(by(the(anthropogenic((increase(of(greenhouse(gases(in(the(atmosphere(and(other(anthropogenic(forces((IPCC(AR5,(Working(Group(1).(

!

!

2!The!historical!dataset!is!public!domain!from!the!Princeton!Land!Surface!Hydrology!Research!Group!and!the!data!used!for!this!analysis!covers!the!period!1950:2008.!!For!climate!projections,!data!up!to!2000!is!used!for!bias!correction1.!Projected!changes!in!climate!are!synthesized!from!a!set!of!bias:corrected!and!spatially!disaggregated!climate!projections!based!on!global!circulation!model!(GCM)!runs!from!the!Coupled!Model!Intercomparison!Project!Phase!5!(CMIP5)!of!the!International!Panel!on!Climate!Change!(IPCC)!archive23.!!Conclusions!about!model!trends!are!drawn!for!two!future!time!periods,!2030:50!and!2080:2100,!under!two!emissions!scenarios!(RCP!4.5!and!RCP!8.5),!defined!by!the!IPCC!(see!Box!2).!!!The!report!is!structured!as!follows:!!First,!previous!work!on!climate!change!analysis!in!Uganda!is!presented.!!This!is!followed!by!the!methodology!section,!where!the!approach!of!the!analysis!is!detailed,!including!issues!related!to!the!the!spatial!and!temporal!resolution.!!The!next!section,!which!is!the!bulk!of!the!report,!provides!analysis!of!historical!climate!and!trends,!as!well!as!future!climate!projections!by!indicator.!!The!main!findings!are!summarized!and!conclusions!are!provided!in!the!final!section.!!!

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!Data!for!bias!correction!are!required!to!be!stationary,!and!therefore!the!period!up!to!2000!is!used.!2!Global!Circulation!Models!(GCMs)!are!three:dimensional!numerical!models!that!use!mathematical!representations!of!the!physical!properties!of!the!atmosphere,!cryosphere,!land!surface,!and!ocean!to!make!projections!about!trends!and!interactions!of!these!processes.!!GCMs!constitute!the!most!advanced!climate!modeling!tools!for!projecting!the!outcomes!of!various!greenhouse!gas!concentrations!in!the!atmosphere!(IPCC,!2013).!3!The!Intergovernmental!Panel!on!Climate!Change!(IPCC)!is!a!scientific!body!under!the!support!of!the!United!Nations!that!reviews,!synthesizes,!and!reports!the!most!recent!scientific!evidence!on!climate!change.!!Part!of!IPCC’s!work!involves!reporting!on!the!simulations!from!a!collection!of!the!most!scientifically!validated!GCMs!based!on!a!set!of!greenhouse!gas!concentration!trajectories.!!The!CMIP5!is!the!latest!set!of!coordinated!climate!model!experiments,!reported!in!the!IPCC’s!Fifth!Assessment!Report!(AR5).!

Box(2:(RCP(scenarios((

The(climate(science(community(continues(to(develop(improved(methods(of(defining(future(scenarios(of(climate(forcing(and(approaches(for(modeling(climate(in(response(to(these(forcings.((Such(activities(have(led(to(the(release(of(an(updated(set(of(global(climate(projections,(labeled(CMIP5((Coupled(Model(Intercomparison(ProjectWPhase(5).((The(IPCC's(Fifth(Assessment(Report,(parts(of(which(have(recently(been(released,(was(based(on(projections(from(CMIP5.((The(CMIP5(uses(new(concentration(trajectories,(referred(to(as(Representative(Concentration(Pathways((RCPs)(that(replace(the(previous(Special(Report(on(Emissions(Scenarios((SRES)(emissions(trajectories.(((

RCPs(are(defined(by(their(radiative(forcing(relative(to(1750(levels((in(watts(per(meter(squared,(W/m2).((Radiative(forcing(describes(the(difference(between(the(amount(of(energy(entering(the(Earth’s(atmosphere(as(sunlight(and(the(amount(of(energy(reflected(back.((A(positive(radiative(forcing(results(in(a(global(warming(and(a(negative(radiative(forcing(results(in(a(global(cooling.((There(are(multiple(natural(and(manWmade(factors(that(affect(radiative(forcing,(including(cloud(coverage,(land(surface(albedo,(volcanic(activity,(and(the(concentration(of(greenhouse(gases(in(the(atmosphere.((A(higher(concentration(of(greenhouse(gases(in(the(atmosphere(reduces(the(amount(of(energy(being(reflected(back(into(space,(resulting(in(a(higher(radiative(forcing.((((

The(IPCC(defines(four(scenarios,(which(describe(four(possible(climate(futures(depending(on(the(amount(of(greenhouse(gases(emitted(in(the(years(to(comeW(RCP(2.6(is(the(lowest(emissions(scenario(because(it(assumes(a(relatively(high(amount(of(mitigation(is(taking(place(throughout(the(century.((RCP(4.5(and(RCP(6(are(two(moderate(emissions(scenarios(that(result(in(climate(‘stabilization’(by(the(end(of(the(century,(meaning(temperatures(are(projected(to(eventually(stabilize(by(the(end(of(the(century.((RCP(8.5(is(the(highest(emissions(scenario(where(mitigation(does(not(take(place(and(temperatures(continue(to(increase(rapidly(throughout(the(century.(((The(difference(in(projected(global(surface(temperatures(as(a(result(of(these(RCP(

scenarios(is(shown(in(the(figure(to(the(left.((Figure(SPM.7(a(from(the(Summary(for(Policy(Makers,(IPCC(AR5(WG(1((

!

!

3!

PREVIOUS(WORK(!This!profile!is!produced!in!the!context!of!various!other!reports!attempting!to!characterize!historical!and!future!climate!in!Uganda,!some!of!which!have!been!incorporated!into!the!recent!!“Integration!of!Climate!Change!Policy!Concerns!into!the!Second!National!Development!Plan!!(NDP!II)”!(NPA,!2014).!!Table"2!summarizes!the!most!recent!reports!on!climate!change!in!Uganda.!!!"Table"2:"Summary!of!Previous!work!Report"Name"

Models" Time"Periods"

Emissions"Scenarios"

Historical"Data"

Downscaling"Technique"

Spatial"Resolution"

DewPoint!(2012)!

10!models!from!CMIP3!

mid!century!!end:of!century!

SRES!A2! Climate!Research!Unit!data!

Statistical!and!dynamical!

One!grid!point!for!all!of!Uganda!~5°x5°!

USAID!(2013)!

10!models!!from!CMIP5!

2015:2045! RCP!4.5!RCP!8.5!

16!Station!data!

Statistical!downscaling:no!bias!correction!

16!zones!across!Uganda!~2°"x!2°!

Baastel!(2014)!

4!models!from!CMIP!5!

mid:century!end:of:century!

RCP!4.5!RCP!8.5!

Model!projections!

Dynamic!Downscaling!

0.44°"x!0.44°!

!!The!analysis!presented!in!this!report!builds!on!previous!efforts!by!providing!a!more!complete!analysis!based!on!a!larger!set!of!GCMs!from!CMIP5,!downscaled!to!a!0.5°"x!0.5°"resolution.!!A!summary!of!the!models!used!in!this!analysis!is!provided!in!Table"3.!!Because!model!assumptions,!scale,!and!techniques!vary!across!GCMs,!the!IPCC!recommends!using!a!multi:model!approach.!!That!is,!they!recommend!drawing!conclusions!about!climate!projections!and!uncertainty!from!all!validated!GCMs,!with!the!underlying!rationale!being!that!all!such!model!projections!must!be!treated!as!‘equally!likely.’!!This!approach!is!adopted!here!and!described!further!below.!!The!report!also!builds!on!previous!work!by!synthesizing!climate!data!into!a!number!of!meaningful!indicators!(refer!above,!Table"1).!!By!using!a!larger!array!of!climate!indicators!–!in!conjunction!with!the!large!set!of!GCMs!–!this!analysis!represents!the!most!comprehensive!such!effort!in!Uganda!to!date.!!!!!Table"3:"Summary!of!CMIP5!GCMs!used!in!this!analysis,!all!downscaled!to!a!0.5°"x!0.5°"resolution!Climate"Scenario" Time"Step" GCM"Combinations"RCP"4.5"" Daily! 9!GCMs!from!1900:2100!

Monthly! 23!GCMs!from!1900:2100!RCP"8.5" Daily! 8!GCMs!from!1900:2100!

Monthly! 20!GCMs!from!1900:2100!!(

(METHODOLOGY!!The!analysis!carried!out!for!this!work!is!based!on!a!multi:model!ensemble,!as!mentioned!above.!!Models!run!on!a!daily!time!step!include!precipitation,!and!minimum,!maximum,!and!mean!daily!temperature,!while!models!run!on!a!monthly!time!step!include!monthly!mean!precipitation!and!monthly!mean!temperature.!Analysis!for!historical!climate!is!performed!using!precipitation,!and!minimum,!maximum!and!mean!temperature!on!a!daily!time!step!from!the!Princeton!dataset!at!a!0.5°"x!0.5°!resolution.!!Because!models!vary!in!spatial!resolution,!characterizing!climates!across!models!first!requires!bias!correction!and!spatial!disaggregation!(BCSD).!!Model!results!and!historical!data!are!spatially!disaggregated!to!a!0.5°"x!0.5°"resolution"for!the!1900:2100!time!period.!!Models!are!then!bias!corrected!to!be!statistically!similar!to!

!

!

4!historical!data!for!the!1950:2000!time!period.!!A!more!detailed!description!of!the!BCSD!approach!used!in!this!analysis!is!provided!in!Annex!C.!!!!Climate!projections!can!be!presented!at!various!resolutions.!!In!this!report,!maps!are!presented!with!data!shown!at!a!0.5°"x!0.5°"resolution.!!However,!climate!indicators!are!reported!spatially!averaged!at!both!the!country:wide!and!the!water!management!zone!(WMZ)!levels,!of!which!there!are!four!in!Uganda:!the!Upper!Nile,!Kyoga,!Victoria!and!Albert!(see!Figure"1).!There!are!various!reasons!to!report!results!in!this!way.!!First,!!climate!and!hydrologic!systems!are!closely!inter:connected;!climate!is!a!driving!force!behind!dynamic!hydrologic!processes!and!conditions.!!Second,!the!natural!hydrologic!unit!is!a!catchment!and!Uganda!is!currently!following!catchment!boundaries!–!the!WMZs!–!for!water!management!and!development,!making!information!at!this!level!of!practical!use.!!Finally,!GCMs!perform!best!at!a!coarser!resolution,!so!it!can!be!misleading!to!report!results!at!a!higher!(e.g.,!0.5°"x!0.5°)"resolution,!even!if!they!are!technically!available!(see!Box!3).!!At!the!same!time,!reporting!only!at!the!country!level!could!fail!to!capture!important!regional!extremes!that!are!muted!out!by!spatial!averaging.!Thus,!results!are!presented!at!both!the!country:wide!and!WMZ!levels,!providing!a!balance!between!accuracy!(or!confidence!in!projections)!and!precision!(or!resolution).!!Victoria!WMZ!is!treated!slightly!differently!from!the!other!three!WMZs!because!it!!consists!of!land!area!and!a!part!of!Lake!Victoria,!which!could!have!quite!different!local!climates.!!Thus,!the!results!of!the!analysis!for!Victoria!WMZ!are!

presented!in!three!groupings:!the!land!portion!of!the!WMZ,!the!land!and!lake!portion!of!the!WMZ,!and!the!larger!Victoria!basin!which!includes!parts!of!Kenya,!Tanzania,!Rwanda!and!Burundi.!!Since!model!results!can!differ!in!both!magnitude!and!direction!of!change,!an!effort!is!made!to!capture!the!full!range!of!model!projections.!!For!each!indicator,!the!median!model!result!and!the!range!of!model!results!is!presented.!!The!level!of!model!agreement!is!also!given!(model!agreement!is!characterized!here!when!at!least!two:thirds!of!the!models!agree!on!the!direction!of!change).!!!!The!mean!across!models!is!not!used!to!communicate!results!in!this!report!as!it!masks!the!range!of!results!across!models!and!can!be!misleading!when!communicating!the!degree!of!model!agreement.!!A!model!‘average’!of!near!zero!could!be!the!result!of!most!models!projecting!near!zero!change,!but!also!the!result!of!two!opposing!changes!that!differ!in!sign,!as!seen!in!Figure"2.!!!

Figure&1:&&Uganda(WMZs.((The(Ugandan(Department(of(Water(Resource(Management(produced(this(map.(

Upper&Nile&

Albert

Box(3.(GCM(ensembles(and(resolution((

GCMs(can(be(run(at(various(spatial(and(temporal(resolutions(and(coarser(GCM(outputs(can(be(downscaled(to(finer(resolutions.((However,(given(the(complexity,(downscaling(to(finer(resolutions(is(quite(computationally(expensive(to(obtain((i.e.(requires(massive(computing(abilities(and(take(a(long(time).((Time(and(cost(constraints(often(do(not(allow(the(use(of(more(than(a(couple(of(GCMs(in(such(exercises.((Given(this(complexity,(GCMs(are(typically(run(on(a(spatial(scale(ranging(from(250(km(to(600(km(wide(grid(cells(and(on(daily(or(monthly(time(steps.((At(this(scale(local(influences,(such(as(smaller(water(surfaces,(topography,(and(clouds,(are(not(fully(captured(in(the(models.((For(this(reason,(models(tend(to(lose(accuracy(at(finer(spatial(resolutions.((That(said,(running(multiple(GCMs(at(a(coarse(resolution(may(provide(more(insight(into(a(range(of(possible(futures(than(the(more(detailed(information(obtained(from((fewer)(finer(resolution(GCMs.((For(more(information(on(downscaling(techniques(see(Annex(D.((

Victoria&

Kyoga&

!

!

5!

It!is!important!to!highlight!that!the!methodological!approach!adopted!for!this!analysis!involved!a!number!of!choices!in!terms!of!the!number!of!models!to!report!on,!the!downscaling!method,!and!the!manner!in!which!results!are!presented,!amongst!others.!!Several!of!these!choices!are!summarized!above.!!Annex!D!provides!a!more!detailed!explanation.!!!!Finally,!a!caveat!is!needed!for!this!work,!and!indeed!all!work!on!climate!projections.!The!science!of!climate!change!is!a!complex!and!inexact!one.!A!number!of!intricate!steps!are!involved!in!projecting!future!climate,!each!associated!with!various!types!of!uncertainties.!Key!ones!arise!from!the!scenarios!of!future!states!of!the!world!and!the!assumptions!surrounding!them,!difficulties!with!modeling!the!complex!climate!system,!lack!of!information!and!data,!and!knowledge!gaps.!As!noted!above,!uncertainties!increase!the!longer!the!time!horizon!and!the!smaller!the!scale.!!That!said,!the!scientific!community!places!considerable!–!and!increasing!–!confidence!in!the!ability!of!climate!models!to!project!future!climate.!This!is!for!a!number!of!reasons!including!that!models!are!built!using!mathematical!representations!of!proven!physical!properties!and!laws!and!that!they!are!increasingly!skillful!at!simulating!past!climate!(perform!well!when!tested!against!observed!climate!data).

Box(4.(GCM(ensemble(range(and(average(((

It(is(currently(impossible(to(attribute(probabilities(to(the(outcome(of(any(GCM.((This(analysis(therefore(assumes(that(each(GCM(projection(is(equally(likely.(Model(projections(can(vary(significantly(as(a(result(of(different(model(assumptions.((Feedback(effects(from(components(such(as(atmospheric(water(vapor,(clouds,(and(snow(melt(all(affect(model(results(and(differ(from(model(to(model.((Therefore,(for(planning(purposes,(

it(is(important(to(consider(what(models(agree(upon(and(the(range(of(uncertainty(across(model(projections,(and(not(simply(report(on(the(average(across(GCMs(or(one(single(GCM(output.((

Figure&2:&&Example(of(how(misleading(is(can(be(to(present(the(median(model(result(alone.((

!

!

6!

HISTORICAL(CLIMATE(AND(TRENDS((Uganda,!being!situated!on!the!equator,!can!be!characterized!as!humid!and!warm!with!high!spatial!and!temporal!variation!in!both!precipitation!and!temperature.!!This!high!level!of!fluctuation!is!a!result!of!variation!in!topography!and!winds,!as!well!as!the!presence!of!large!lake!surface!areas!and!river!areas.!In!Box!5!and!Box!6!that!follow,!Uganda’s!temperature!and!precipitation!are!characterized!respectively!by!considering!spatial!and!seasonal!variation,!inter:annual!variability,!and!trends!since!1950.!!

Historical!temperatures!in!Uganda!average!22°C!with!maximum!daily!temperatures!averaging!29°C!and!minimum!daily!temperatures!averaging!17°C.!!!!!As!can!be!seen!in!Figure"3,!temperatures*vary*spatially*across!Uganda!with!the!highest!temperatures!occurring!in!the!Northern!part!of!the!country!and!lowest!temperatures!occurring!in!the!southeastern!and!southwestern!sides.!!!The!Upper!Nile!WMZ!is!the!warmest!WMZ!with!a!mean!annual!temperature!of!23.8°C.!!The!Victoria!WMZ!is!the!coolest!WMZ!with!a!mean!annual!temperature!of!21.0°C.!!Figure"3"also!shows!that!temperatures*vary*seasonally.!!December:February!is!the!warmest!season!of!the!year!and!June:August!is!the!coolest!season!of!year.!!! Figure&3:(Seasonal(mean(temperature(

Figure&4:(Box(plots(of(daily(temperature(by(month(

Figure&5:(Month(mean(temperature(from(1950W2008(

Figure"4"is!a!plot!of!the!range!of!historical!mean!daily!temperatures!by!month.!!While!there!is!a!clear!seasonality!to!daily!temperature,!the!large!majority!of!daily!temperatures!stay!within!2°C!of!the!annual!mean.!!!This!figure!also!shows!that!there!is!variability!within!months,!with!the!warmer!months!having!higher!inter:monthly!variability!than!colder!months.!!However,!daily!temperatures,!for!the!most!part!stay!within!1°C!of!the!median!monthly!value.!!!Figure"5!is!a!plot!of!historical!mean!monthly!temperature!from!1950:2008.!!This!plot!shows!that!most!mean!monthly!temperatures!stay!within!1.5°C!of!the!mean.!!!There!is!also!a!noticeable!inter9annual*variability.!!!For!example,!in!the!1960s!there!were!two!years!in!a!row!where!mean!monthly!temperatures!stayed!below!23°C.!!!Climate(Trends(Temperatures*have*been*increasing!in!Uganda!by!approximately!0.2°C/decade!over!the!last!30!years!(DEWPoint!2012).!Figure"5"confirms!a!clear!trend!in!increasing!temperature,!especially!noticeable!in!the!last!decade.!!!!

Box(5:(Historical(Temperature.(

!

!

7!

(

Figure&8:(Monthly(mean(precipitation(from(1950W2008(

Figure&7:(Box(plots(of(monthly(precipitation(by(month(

Figure&6:(Seasonal(monthly(precipitation(

Annual!rainfall!in!Uganda!is!on!average!1180!mm!and!ranges!from!1135!mm!to!1264!mm.!!!!Rainfall*varies*spatially!across!Uganda!with!the!southern!part!of!the!country!receiving!annual!rainfall!in!excess!of!2100!mm/year!and!the!northeastern!part!of!the!country!(the!driest!region)!receiving!less!than!500mm/year!(see!Figure"6).!!!Uganda’s!rainfall!patterns!can!be!explained!in!part!by!sea!surface!temperature!in!the!equatorial!Pacific!Ocean!the!Indian!Ocean!(Tetra!Tech,!2013).!!The!Victoria!WMZ!receives!the!highest!rainfall!of!the!four!WMZs!and!the!Upper!Nile!WMZ!receives!the!lowest!rainfall!levels.!!!Figure"6"also!shows!that!rainfall*varies*seasonally!in!Uganda!with!December:February!being!the!driest!season!and!March:May!being!the!wettest!season.!!

Figure"7"is!a!plot!of!monthly!precipitation!by!month.!!As!can!be!seen!in!this!figure,!rainfall!in!Uganda!follows!a!bimodal!pattern.!!!The!long!rains!occur!from!March!to!May!and!the!short!rains!occur!from!September!to!October.!!The!northwestern!part!of!the!country,!however,!sees!only!one!long!rainy!season!from!March!to!November.!!Note!also!that!the!dry!seasons!correspond!to!the!warmer!seasons!in!Uganda.!!!Figure"8"is!a!plot!of!monthly!precipitation!from!1950:2008,!showing!that!there!is!inter9annual*variability*in!rainfall.!!!!The!two!stars!in!Figure"8!are!1962!and!1997!monthly!precipitation!highs!and!mark!the!two!most!extreme!rainfall!events!on!record.!!These!were!also!the!two!years!in!this!time:frame!when!high!levels!of!flooding!were!experienced:!!High!rainfall!in!the!early!1960s!led!to!increased!water!levels!in!Lake!Victoria.!!These!levels!remained!high!until!2005,!when!lake!levels!fell!again,!in!part!due!to!low!rainfall.!Rainfall!extremes!in!1997/98!increased!levels!in!Lake!Kyoga!by!over!2m,!flooding!an!area!of!approximately!580!km2.!!!!!Climate(Trends(Recent!studies!have!found!that!there!has!been!a!decrease*in*rainfall*during*the*MAM*rainy*season!(Baastel,!2014),!though*no*statistically*significant*changes*in*annual*rainfall*have!been!found.!!!

Box(6:(Historical(Precipitation.(

*!

*!

!

!

8!

CLIMATE(CHANGE(PROJECTIONS(OVERVIEW(!This!section!presents!GCM!projections!for!climate!in!Uganda,!first!by!providing!broad!changes!in!projected!temperature!and!precipitation!and!then!by!describing!each!indicator!in!detail.!!Temperature(Change(As!previously!mentioned,!there!has!already!been!a!noticeable!warming!taking!place!in!Uganda.!!Temperatures!are!approximately!0.6!°C!higher!on!average!than!they!were!three!decades!ago.!!!Figure"9!is!a!plot!of!historical!and!future!model!projections!of!mean!annual!temperature.!!!The!shaded!region!captures!the!range!of!uncertainty!across!all!models!run!under!the!same!emissions!scenario.!!RCP!4.5,!the!blue!band,!is!the!‘moderate’!emissions!scenario!and!RCP!8.5,!the!red!band,!is!the!high!emissions!scenario.!!!As!the!figure!shows,!the!moderate!emissions!scenario!is!projected!to!lead!to!temperature!stabilization!by!the!end!of!the!century!at!about!2°C!warmer!than!historical!values.!!The!higher!emissions!scenario,!however,!projects!a!continued!increase!in!temperature!throughout!the!century,!reaching!temperatures!4:6°C!warmer!than!historical!values!by!the!end!of!the!century.!!!

!Figure& 9:& Historical( and( future( annual( temperature,( spatially( averaged( across( Uganda.( ( Bold( lines(represent(the(model(median(projection(for(each(year(and(the(shaded(bands(span(the(range(of(the(model(projections(for(each(year.((The(yearly(median(projection(is(smoothed(out(by(taking(a(threeWyear(moving(average.((

(

(

Temperature*Variability**Figure"10!shows!that!all!models!project!an!increase!in!annual!temperature!and!that!the!range!of!projections!is!increasing!towards!the!middle!of!the!century.*

Figure&10:(GCM(output(for(Mean(Annual(Temperature(of(multiple(models(from(2000W2050((

!

!

9!

((Precipitation(Change((There!has!been!no!discernible!change!in!annual!rainfall!in!Uganda!beyond!natural!variability.!!!Some!recent!studies!suggest!a!decrease!in!rainfall!during!the!long!rains!(March:May),!though!these!are!the!first!to!note!any!such!change!(Baastel,!2014).!!!Figure"11!is!a!plot!of!historical!and!future!model!projections!of!mean!annual!precipitation.!!The!shaded!bands!capture!the!range!of!uncertainty!across!all!models!run!under!the!same!RCP!scenario.!!RCP!4.5!is!the!‘moderate’!emissions!scenario!and!RCP!8.5!is!the!high!emissions!scenario.!!!Based!on!Figure"11,!there!is!no!clear!trend!on!direction!of!change,!though!it!seems!to!be!increasing!towards!the!end!of!the!century.!!There!is!a!discernible!trend,!however,!that!the!range!of!annual!precipitation!values!within!each!run!is!increasing!throughout!the!century!(see!Figure"12),!with!some!models!projecting!values!well!beyond!the!envelope!of!historical!levels,!both!higher!and!lower.!!!

!Figure& 11:& Historical( and( future( annual( precipitation( spatially( averaged( over( Uganda.( ( Bold( lines(represent( model( median( projections( for( each( year.( The( shaded( bands( span( the( range( of( model(projections( for( each( year.( ( The( yearly( median( projection( is( smoothed( out( by( taking( a( threeWyear(moving(average.(

!*

Figure&12:(GCM(output(for(Mean(Annual(Precipitation(of(multiple(models(from(2000W2050((each(line(represents(the(output(of(one(model(run).(

Precipitation*Variability**Figure"12!shows!that!models!project!an!increase!in!the!range!of!annual!precipitation!though!there!is!disagreement!amongst!models!on!the!direction!of!change.!*

!

!

10!

CLIMATE(PROJECTIONS(BY(INDICATOR(!Interpreting(Graphics(and(Data(This!section!provides!the!projections!for!each!indicator!in!detail!(one!indicator!per!page).!Several!types!of!graphics!are!used!to!present!the!results,!and!the!table!below!explains!how!each!of!these!should!be!interpreted.!!!Annex!A!shows!the!derivation!of!the!more!complex!indicators.!The!coefficient!of!variation!(CV)!–!the!standard!deviation!divided!by!the!mean!–!has!been!calculated!for!a!number!of!indicators.!!Annex!B!includes!additional!CV!calculations!that!are!not!included!in!this!section.

Bar"Charts"with"error"

The!horizontal!axis!defines!the!range!of!values,!so!in!this!example!‘30’!represents!the!range!30–50,!‘50’!represents!the!range!of!50–70,!etc.!!The!vertical!axis,!unless!otherwise!specified,!is!the!annual!frequency!corresponding!to!that!range.!!Bars!for!these!plots!are!the!median!model!projection!for!the!corresponding!range!in!each!time!period.!The!error!bars!show!the!frequency!range!of!model!projections!for!that!time!period!and!emissions!scenario.!!For!example,!by!mid:century,!the!median!model!projects!values!between!50!and!70!to!occur!about!once!every!10!years!(i.e.!0.1).!!However,!some!models!project!this!frequency!to!be!as!high!as!0.45!and!as!low!as!zero.!!!

Maps"

In!maps!such!as!this,!the!colors!indicate!a!range!of!projected!changes!where!at!least!66%!of!the!models!agree!on!the!direction!of!change!(or!agree!on!no!change).!!Grey!areas!mean!that!fewer!than!66%!of!the!models!agree!on!the!direction!of!change.!!The!color!shows!the!median!model!projection.!!

Box"plots"

Two!types!of!box!plots!are!used!to!present!results.!!Data!are!either!grouped!by!month,!such!as!the!top!plot!here,!where!data!from!all!models!within!specified!emissions!scenario!and!time!period!are!plotted!in!one!box!per!month,!OR,!data!are!grouped!by!model,!such!as!the!bottom!plot!here,!where!each!box!represents!data!from!one!model!(so!9!models!are!shown!in!this!plot).!!!Each!box!in!a!plot!represents!the!interquartile!range!of!the!data,!which!is!the!middle!50%!of!the!data!in!that!grouping.!!The!red!line!in!the!box!is!the!median!value!of!that!grouping.!!The!lines!extending!vertically!out!from!the!box!(i.e.!the!whiskers)!are!at!most!1.5!times!the!length!of!the!box,!or!the!furthest!point!from!the!box,!whichever!is!shorter.!!The!red!crosses!beyond!the!whiskers!are!the!outliers.!!!The!dotted!lines!included!across!the!plots!are!the!interquartile!range!of!historical!values!and!the!solid!line!is!the!historical!median,!unless!otherwise!specified.!!They!are!intended!as!reference!points.!!!

!!mod.!!mod.!!mod.!!mod.!!mod.!!mod!.!mod.!!mod.!mod.!!!!!1!!!!!!!!!!2! !!!!!!!3!!!!!!!!!4!!!!!!!!!!!5!!!!!!!!6!!!!!!!!!!7!!!!!!!!!8!!!!!!!!!9!

! 11!MEAN%TEMPERATURE%

Shade&=&at%least%90%%of%models%project%an%increase%in%Mean%Temperature&

Figure&12:&Boxplots%of%mean%monthly%temperature,%including%all%monthly%projections%for%specified%emissions%scenarios%and%time%periods.%The%blue%line%is%the%historical%mean%temperature%and%black%dotted%lines%are%the%historical%interquartile%range.&

Figure&11:&Median%projected%change%in%average%daily%temperature%(°C)%from%historical%baseline.%%&

Mean%Temperature:%%Daily!averaged!temperature!in!°C.!!!!!!

• All#models#project!an!increase!in!mean!temperature!for!both!emissions!scenarios!and!both!time!periods%

• By!mid=century,!models!project!an!increase!in!mean!annual!temperature!between!1.2!and!2.3°C!averaged!across!Uganda.!This!is!projected!to!increase!to!between!1.7!to!5.6°C!by!end!of!century.!!The!high!emissions!scenario!is!projected!to!lead!to!a!higher!change!in!temperature!with!median!models!projecting!a!difference!of!0.4°C!by!mid!century!and!1.9°C!by!end!of!century!between!emissions!scenarios!(Table&4).%

• As!shown!in!Figure&11,!warming#is#projected#to#vary#spatially!across!Uganda!with!the!central!region!of!the!country!warming!more!than!the!northern!and!Lake!Victoria!region.%

• Monthly!mean!temperatures!are!projected!to!exceed!historical!monthly!temperatures!for!all#months!(Figure&12).!!Further,!future!monthly!projections!fall!beyond!historical!normal!ranges!(i.e.!the!interquartile!range).!%

Table&4:&Mean%annual%temperatures%across%Uganda%and%by%WMZ%

&&Region&

&1950:2000&

2030:2050& 2080:2100&Mod.&

emissions&High&

emissions&Moderate&emissions&

High&emissions&

Uganda&22.7&

24.1&(23.9P24.7)%

24.5&(23.8P25.0)%

24.9&(24.4P25.8)%

26.8&(26.0P28.3)%

Albert&22.1&

23.7&(23.4P24.0)%

23.9&(23.5P24.5)%

24.5&(23.8P25.2)%

26.3&(25.6P27.8)%

Kyoga&23.2&

24.6&(24.2P25.2)%

24.9&(24.3P25.5)%

25.3&(24.6P26.5)%

27.6%(26.2P28.9)%

Upper&Nile& 23.8&

25.2&(24.8P25.8)%

25.5&(24.4P26.1)%

26.0&(25.4P27.2)%

27.5&(27.2P29.4)%

Victoria&Land& 21.1&

22.5&(22.4P22.9)%

22.9&(22.6P23.4)%

23.4&(22.8P24.0)%

25.5&(24.6P26.5)%

VictoriaWMZ& 21.6&

23.0&(22.9P23.5)%

23.4&(23.0P23.9)%

23.8&(23.4P24.6)%

26.0&(25.0P27.1)%

Victoria&Basin& 21.0&

22.4&(22.3P23.0)%

22.8&(22.4P23.3)%

23.2&(22.8P24.0)%

25.4&(24.4P26.6)%

!

! 12!

%&&Region&

1950&:&

2000&

2030:2050& 2080:2100&Moderate&emissions&

High&emissions&

Moderate&emissions&

High&emissions&

Uganda&29.6&

31.6&(31.3P32.2)%

32.0&(31.2P32.5)%

32.4&(31.9P33.5)%

34.4&(33.6P35.9)%

Albert&28.6&

30.9&(30.6P31.3)%

31.1&(30.7P31.8)%

31.8&(31.0P32.5)%

33.6&(33.0P35.1)%

Kyoga&30.4&

32.5&(32.0P33.2)%

32.8&(32.0P33.5)%

33.3&(32.4P34.5)%

35.7&(34.1P36.9)%

Upper&Nile& 30.9&

32.7&(32.4P33.5)%

33.2&(31.8P33.8)%

33.7&(33.0P34.9)%

35.3&(34.9P37.2)%


29.8&(29.7P30.3)%

30.2&(29.9P30.8)%

30.7&(30.1P31.4)%

32.9&(32.0P33.9)%

Victoria&Uganda& 28.4&

30.3&(30.2P30.9)%

30.8&(30.3P31.2)%

31.1&(30.7P32.0)%

33.4&(32.4P34.5)%


29.8&(29.7P30.4)%

30.2&(29.7P30.7)%

30.6&(30.2P31.5)%

32.9&(31.9P34.1)%

!

MAXIMUM%DAILY%TEMPERATURE%

Figure&15:&Median%model%projections%of%annual%number%of%days%that%maximum%temperature%>%30°C.%%

Figure&13:&Annual%

frequency%distribution%of%maximum%daily%temperature.%%The%band%around%each%bar%shows%the%range%of%model%projections.!!

Figure&14:&Increase%in%maximum%daily%

temperature%from%historical%averages.%

Maximum%Daily%Temperature:%Daily!temperature!highs!in!°C.!!!!!!

• All#models#project!an!increase!in!maximum!daily!temperature!for!both!emissions!scenarios!and!both!time!periods!with!an!increase!between!1.6°C!and!2.9°C!by!mid!century!and!between!2.3°C!and!6.3°C!by!end!of!century!(Table&5).%

• Figure&13&shows!that!the!distribution!in!maximum!daily!temperature!is!projected!to!shift!up.%

• As!shown!in!Figure&14,!increase!in!maximum!temperature!is!projected!to!vary#spatially!across!Uganda!with!the!central!region!of!the!country!warming!more!than!the!northern!and!Lake!Victoria!region.!%

• Figure&15&shows!the!spatial!variation!of!number!of!days!a!year!with!temperatures!exceeding!30°C.!!Historically!temperatures!exceed!30°C!in!the!northern!part!of!the!country.!!With!the!exception!of!the!southeastern!and!southwestern!sides,!the!number!of!days!exceeding!30°C!is!projected!to!increase!dramatically!throughout!the!country.%

Table&5:&Average%maximum%daily%temperature%

Shade&=&>90%%of%models%project%an%increase%in%maximum%daily%temperature&

! 13!MINIMUM%DAILY%TEMPERATURE%

Figure&16:&Annual%

frequency%distribution%of%minimum%daily%temperature%%!!!

Figure&17:&Increase%in%

minimum%daily%temperature%from%historical%averages.% Figure&18:&Median%

model%projections%of%minimum%daily%temperature.%%%

Minimum%Daily%Temperature:%Daily!temperature!lows!in!°C.!!!!!!

• All#models#project!minimum!daily!temperature!to!increase!for!both!emissions!scenarios!and!both!time!periods!with!an!increase!between!0.8°C!and!1.8°C!by!mid!century!and!between!1.3°C!and!4.3°C!by!end!of!century!(Table&6).%

• Figure&16&shows!how!the!distribution!in!maximum!daily!temperature!is!projected!to!shift!up.%

• Figure&17!shows!how!the!increase!in!minimum!daily!temperature!across!emissions!scenarios!and!time!periods!varies!spatially.!!Note!that!minimum!daily!temperature!is!projected!to!increase!less!than!mean!or!maximum!daily!temperature.!!!%

• Figure&18&shows!the!spatial!variation!of!minimum!daily!temperature.!!It!can!be!seen!that!historically,!minimum!daily!temperature!remains!below!20°C!for!the!whole!region.!!The!central!and!northern!part!of!the!country!is!projected!to!see!higher!minimum!daily!temperatures!whereas!the!southeastern!and!southwestern!sides!are!projected!to!remain!the!cooler!parts!of!the!region.%

Table&6:&Average%minimum%daily%temperature%&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& 17.1&

18.1&(17.9P18.6)%

18.4&(17.9P18.9)%

18.8&(18.4P19.8)%

20.7&(19.9P22.2)%

Albert&16.8&

17.9&(17.7P18.2)%

18.1&(17.7P18.7)%

18.7&(18.0P19.3)%

20.4&(19.8P21.9)%

Kyoga&17.4&

18.5&(18.1P19.0)%

18.7&(18.1P19.2)%

19.0&(18.5P20.2)%

21.3&(19.9P21.9)%

Upper&Nile& 18.0&

19.1&(18.8P19.6)%

19.4&(18.5P19.9)%

19.7&(19.3P20.9)%

21.3&(20.9P23.1)%


16.6&(16.6P17.1)%

17.0&(16.8P7.5)%

17.5&(16.9P18.1)%

19.6&(18.7P20.6)%


17.1&(17.1P17.6)%

17.5&(17.1P17.9)%

17.9&(17.5P18.7)%

20.1&(19.1P21.2)%


16.6&(16.5P17.1)%

16.9&(16.5P17.4)%

17.3&(17.0P18.1)%

19.5&(18.5P20.7)%

!Shade&=&>%90%%of%models%project%an%increase%in%minimum%daily%temperature&

! 14!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& 0.0&

0.3&(0.1P0.9)%

0.6&(0P3.3)%

1.3&(0.8P5.7)%

18.3&(5.9P71.1)%

Albert&0.0&

0.3&(0P0.8)%

0.3&(0P3.5)%

1.2&(0.2P4.7)%

11.4&(4.6P80.0)%

Kyoga&0.0&

0.3&(0.1P0.8)%

0.6&(0.1P3.4)%

1.3&(0.5P6.5)%

31.5&(4.8P74.7)%

Upper&Nile& 0.0&

0.8&(0.2P1.8)%

1.6&(0.1P5.7)%

3.0&(1.2P9.0)%

12.3&(10.1P54.4)%

Victoria&Land& 0.0&

0.0&(0P0.2)%

0.0&(0P1.0)%

0.4&(0P2.4)%

17.2&(4.2P70.2)%


0.0&(0P0.2)%

0.0&(0P0.8)%

0.3&(0P2.7)%

18.3&(3.9P75.9)%


0.1&(0P0.3)%

0.2&(0P0.5)%

0.6&(0.1P3.2)%

18.7&(5.3P83.7)%

!

HEAT%WAVE%DURATION%INDEX%

Shade&=&at%least%90%%of%models%project%an%increase%in%HWDI&

Figure&19:&Median%model%projections%for%change%in%HWDI%between%projected%and%historical%values.%%

Figure&20:&Median%model%projections%of%HWDI.%%Each%model’s%yearly%HWDI%values%are%captured%in%each%box.%%%

Table&7:&Average%HWDI%%

Heat%Wave%Duration%Index%(HWDI):%The!average!number!of!consecutive!days!a!year!where!daily!temperature!is!at!least!5°C!higher!than!historical!mean!for!that!location.!!

• Historically!there!have!been!nearly!no!heat!waves!in!Uganda,!i.e.!temperatures!very!rarely!exceed!5°C!above!normal!daily!temperatures.!

• All#models#project!an!increase!in!HWDI!between!0.1!and!3.3!days/year!by!mid=century!and!between!0.8!and!71.1!days/year!by!end!of!century!with!models!projecting!Kyoga!WMZ!to!have!higher!HWDI!and!Albert!and!Upper!Nile!WMZs!having!lower!HWDI!(Table&7).!

• Figure&19&shows!the!spatial!variation!of!HWDI!for!both!emissions!scenarios!and!time!periods.!!It!can!be!seen!that!the!northwestern!part!of!the!region!is!projected!to!have!longer!heat!waves!than!the!rest!of!the!country!and!that!the!end!of!century,!high!emissions!scenario!projected!heat!waves!to!be!much!longer!than!the!other!scenarios!and!time!periods.!

• Figure&20&shows!that!within!each!scenario!and!time!period!there!is!not!only!high!variation!across!model!median!projections!of!HWDI!but!there!is!also!high!variation!in!inter=annual!variability!across!models.!!

! 15!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& 10.8&

11.4&(9.9P12.1)%

11.4&(9.8P13.0)%

11.5&(9.7P13.2)%

12.6&(10.3P15.3)%

Albert&9.0&

9.5&(8.2P9.9)%

9.6&(8.2P10.5)%

9.7&(7.9P11.1)%

10.6&(8.6P12.5)%

Kyoga&11.6&

12.2&(11.1P13.0)%

12.2&(10.5P13.6)%

12.2&(11.0P14.1)%

13.6&(11.8P15.9)%

Upper&Nile& 13.1&

14.1&(11.3P15.4)%

13.8&(11.2P17.0)%

13.9&(10.8P16.8)%

14.7&(11.9P19.9)%

Victoria&Land& 8.4&

8.9&(7.8P9.3)%

8.9&(8.1P9.9)%

9.1&(7.7P10.5)%

10.3&(8.0P11.9)%


9.7&(8.7P10.1)%

9.8&(9.1P11.0)%

10.0&(8.6P11.2)%

11.2&(8.8P12.9)%


9.4&(8.6P9.9)%

9.6&(9.1P10.7)%

9.8&(8.6P11.0)%

11.0&(8.6P12.7)%

!

SIMPLE%DAILY%PRECIPITATION%INDEX%

Shade&=&at%least%90%%of%models%project%an%increase%in%sdii%Light&Shade&=&at%least%66%%of%models%project%an%increase%in%sdii%

Figure&21:&Boxplots%of%sdii.%%Each%model’s%projections%are%captured%in%each%box.%%The%three%dotted%horizontal%lines%are%historical%sdii%median%with%25th%and%75th%percentile%on%either%side.%

Figure&22:&Maps%of%historical%and%median%model%projections%of%sdii%

Simple%Daily%Precipitation%Index%(sdii):%The!average!amount!of!rainfall!during!a!wet!day,!where!a!wet!day!means!rainfall!exceeds!1mm/day.!!!!• Most#models!project!there!will!be!an!increase!in!sdii!with!the!median!model!projecting!an!increase!of!0.6mm/wet!day!by!mid=century!and!between!0.7!and!1.8!mm/wet!day!by!the!end!of!the!century!(Table&8).!

• As!is!shown!in!Figure&21,!there!is!a!high#degree#of#variability!among!models!in!terms!of!the!median!sdii!values!as!well!as!noticeable!variability!among!models!in!terms!of!the!range!of!interannual!sdii!values!While!some!models!project!a!range!of!yearly!sdii!to!span!4!mm/wet!day,!other!project!that!range!to!be!greater!than!8!mm/wet!day.!!

• As!is!shown!in!Figure&22,&sdii!varies!spatially!across!Uganda!with!the!northern!part!of!the!country!receiving!sdii!values!as!high!as!20!mm/wet!day!and!the!southern!part!of!the!country!receiving!sdii!values!between!5!and!10mm/wet!day.!!

Table&8:&Average%sdii%in%mm/wet%day%

! 16!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& 111.2&

157.2&(117P222)%

154.2&(119P222)%

158.4&(126P216)%

196.4&(150P255)%

Albert&92.1&

125.4&(95.9P171)%

115.6&(92.5P190)%

131.2&(97.1P182)%

151.8&(119P225)%

Kyoga&119.4&

169.7&(127P254)%

172.5&(130P237)%

187.7&(136P247)%

216.0&(162P278)%

Upper&Nile& 134.3&

199.8&(137P288)%

187.9&(126P297)%

188.9&(147P250)%

248.4&(168P293)%


123.9&(99.3P159)%

119.3&(94.9P164)%

127.3&(99.5P181)%

145.8&(112P219)%


138.6&(110P176)%

140.0&(107P173)%

147.0&(109P194)%

168.7&(127P247)%


136.3&(116P171)%

141.9&(121P170)%

140.3&(118P186)%

174.6&(134P237)%

!

FIVEPDAY%PRECIPITATION%

Shade&=&at%least%90%%of%models%project%an%increase%in%5Pday%precipitation%

Five?Day%Precipitation:!The!highest!annual!precipitation!value!over!a!5=day!period.!!!!!

• All#models#project!an!increase!in!five=day!precipitation!across!Uganda.!!Five=day!precipitation!is!projected!to!increase!by!between!5!and!111!mm!by!mid!century!and!by!15!and!144mm!by!the!end!of!century,!the!high!end!of!these!projections!would!result!in!a!doubling!in!5=day!precipitation!(Table&9).!

• Historically!5=day!precipitation!is!centered!around!100!mm!with!60%!of!the!years!seeing!values!between!100=120mm.!!Models!project!the!distribution!of!5=day!precipitation!to!flatten!out!with!greater!inter=annual!variability!and!a!higher!frequency!of!high!extreme!events!(Figure&23).!

• As!is!shown!in!Figure&24,&5=day!precipitation!varies!spatially.!!Historically!the!northern!region!of!the!country!has!had!higher!values!than!the!southern!region.!!For!both!historical!and!future!projections,!Lake!Victoria!has!a!higher!5=day!precipitation!than!the!regions!immediately!surrounding!the!lake.!!While!all!of!Uganda!is!projected!to!see!an!increase!in!5=day!precipitation,!average!values!are!expected!to!be!as!high!as!300!mm/5=days!in!the!northern!part!of!the!country.!!!!

Figure&23:&Annual%frequency%distribution%of%5Pday%precipitation%extreme.%%Each%bar%is%the%median%model%projection%for%the%specified%time%period%with%the%error%lines%representing%the%range%in%frequency%across%models%for%the%specified%range.%%%

Figure&24:&Historical%and%median%model%projections%of%average%annual%5Pday%precipitation%extremes%in%mm.%%%

>!

>!Table&9:&Average%annual%5Pday%precipitation%extreme%

! 17!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& 24.4&

31.8&(28.8P42.6)%

32.3&(30.3P46.0)%

32.0&(27.1P50.1)%

29.7&(26.0P46.3)%

Albert&20.5&

28.8&(26.0P37.2)%

28.8&(25.2P41.7)%

28.0&(25.6P47.7)%

26.2&(21.1P43.7)%

Kyoga&26.9&

36.1&(28.4P43.9)%

33.9&(31.6P48.5)%

35.3&(25.3P49.7)%

32.3&(28.1P46.5)%

Upper&Nile& 32.0&

39.3&(33.5P50.7)%

40.3&(32.8P51.0)%

41.3&(30.4P60.8)%

35.3&(30.9P53.5)%


31.2&(23.7P36.4)%

32.1&(23.4P41.1)%

28.0&(22.8P42.8)%

26.9&(22.6P45.2)%


28.3&(23.1P38.5)%

30.9&(23.0P42.7)%

27.3&(23.0P42.4)%

30.1&(21.9P46.9)%


40.4&(31.5P54.1)%

44.3&(34.4P63.8)%

41.6&(32.7P57.1)%

43.5&(28.8P67.9)%

!

CONSECUTIVE%DRY%DAYS%

Figure&26:&Median%model%projections%of%change%in%average%number%of%consecutive%dry%days%a%year.%%%%%

Figure&25:&Annual%frequency%distribution%of%consecutive%dry%days.%%Each%bar%represents%the%median%model%projected%frequency%for%that%range%of%values%(here%each%value%is%the%lowest%of%a%10%day%range%of%values)%and%the%error%lines%represent%the%model%range%of%projections%for%that%value.!!!

Shade%=%>90%%of%models%project%an%increase%in%number%of%consec.%dry%days%Shade%=%>66%%of%models%project%an%increase%in%number%of%consec.%dry%days%

Consecutive%Dry%Days:!The!maximum!number!of!consecutive!dry!days!a!year!where!precipitation!is!less!than!1!mm/day.!!!

• All#models#project!an!increase!in!the!number!of!consecutive!dry!days!a!year!Uganda=wide!(Table&10).!!

• Figure&25&shows!that!historically!most!years!have!between!10!and!40!consecutive!dry!days!with!80%!of!the!years!being!between!20!and!30!days.!!The!median!model!projects!that!most!years!will!see!at!least!30!consecutive!dry!days!with!some!years!seeing!as!many!as!60!consecutive!dry!days!by!mid!century!and!80!on!occasion!by!the!end!of!the!century.!!

&!

• Figure&26!shows!that!the!change!in!consecutive!dry!days!is!anticipated!to!vary#spatially!with!the!southern!region!of!the!country!seeing!the!biggest!increase!in!consecutive!dry!days!and!the!northern!part!of!the!country!and!some!parts!of!Lake!Victoria!seeing!the!lowest!increase!in!consecutive!dry!days.!!!

Table&10:&Average%maximum%number%of%consecutive%dry%days%a%year%

! 18!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&&

103.9&&

103.2&(92.0P121)%

103.4&(90.9P126)%

106.4&(90.6P131)%

119.2&(91.3P146)%

Albert& 102.3&&

99.7&(92.8P117)%

100.4&(89.1P119)%

102.2&(94.0P123)%

115.3&(88.0P138)%

Kyoga& 100.4&&

106.0&(91.5P122)%

102.8&(88.2P123)%

109.1&(92.3P136)%

122.4&(94.0P144)%

Upper&Nile&

99.4&&

97.5&(88.6P117)%

99.6&(73.5P123)%

101.4&(90.5P125)%

108.5&(91.5P142)%

Victoria&Land&

106.1&&

105.5&(94.0P124)%

104.3&(91.8P131)%

108.1&(91.2P136)%

120.0&(90.6P154)%

Victoria&Uganda&

113.1&&

113.2&(85.9P135)%

115.2&(87.9P138)%

115.9&(79.4P141)%

130.0&(71.8P161)%

Victoria&Basin&

102.6&&

103.0&(86.2P124)%

103.8&(85.6P130)%

106.0&(84.0P132)%

121.3&(83.8P153)%

!

MONTHLY%PRECIPITATION%

Shade&=&at%least%66%%of%models%project%an%increase%in%monthly%precipitation%&

Figure&28:&Maps%where%>66%%of%models%agree%on%direction%of%change%in%precip%(mm/month).%Grey%means%<66%%of%models%agree.%

Figure&27:&Boxplots%of%monthly%precipitation.%%Each%model’s%range%of%monthly%precipitation%values%is%captured%in%each%box.%%The%two%horizontal%lines%are%the%1962%and%1997%monthly%precipitation%high,%the%two%highest%monthly%values%in%the%past%50%years.%%

Monthly%Precipitation:!the!amount!of!rainfall!per!month!in!mm/month.!!!!

• There!is!uncertainty!among!models!regarding!the!direction!of!change!of!average!monthly!precipitation!by!mid=century.!!Under!the!high!emissions!scenario,!most#models!project!an!increase!in!average!monthly!precipitation!by!the!end!of!the!century!(Table&11).!

• As!shown!in!Figure&27,!there!is!no!remarkable!change!in!model!median!projections!of!monthly!precipitation.!!Most!models,!however,!project!that!there!will!be!an!increase!in!the!extreme!outliers.!!This!figure!shows!that!most!models!project!that!the!monthly!precipitation!levels!of!1962!and!1997!(the!two!highest!on!record)!will!be!exceeded!multiple!times!in!both!20=year!periods!(i.e.!both!mid=century!and!end=of=century!periods).!!

• Figure&28&shows!that!there!is!uncertainty!among!models!regarding!the!direction!of!change!of!precipitation!by!mid=century.!!By!the!end!of!century,!however,!there!is!greater!agreement!among!models!that!the!eastern!half!of!the!country!will!see!an!increase!in!precipitation.!!

Table&11:&Average%monthly%precipitation%%

! 19!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&wide& 1.17&

1.20&(.78P1.86)%

1.22&(.72P2.1)%

1.19&(.85P2.23)%

1.79&(.79P2.99)%

Albert&0.84&

0.85&(.35P1.57)%

0.93&(.35P1.54)%

0.86&(.54P1.82)%

1.49&(.42P2.59)%

Kyoga&1.22&

1.43&(.75P2.07)%

1.34&(.80P2.28)%

1.40&(.90P2.67)%

1.97&(.95P2.85)%

Upper&Nile& 1.17&

1.01&(.58P1.79)%

1.18&(.30P2.37)%

1.24&(.59P2.39)%

1.48&(.93P2.92)%


1.09&(.66P1.93)%

1.16&(.58P1.93)%

1.17&(.72P2.45)%

1.88&(.59P3.24)%


1.45&(.68P2.35)%

1.56&(.68P2.24)%

1.65&(.52P2.94)%

2.23&(.40P3.62)%


1.35&(.78P2.08)%

1.39&(.79P2.25)%

1.52&(.84P2.57)%

2.06&(.75P3.33)%

!&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& 0.15&

0.20&(.08P.37)%

0.22&(.08P.47)%

0.25&(.11P.64)%

0.40&(.15P1.05)%

Albert&0.06&

0.09&(.01P.23)%

0.13&(.02P.29)%

0.10&(.02P.42)%

0.19&(.02P.67)%

Kyoga&0.18&

0.27&(.08P0.43)%

0.24&(.11P0.56)%

0.32&(.15P.72)%

0.47&(.21P1.21)%

Upper&Nile& 0.14&

0.18&(.07P.34)%

0.20&(.02P.49)%

0.20&(.09P.36)%

0.29&(.10P.86)%


0.19&(.04P.47)%

0.21&(.06P0.60)%

0.24&(.06P.89)%

0.45&(.08P1.31)%


0.25&(.06P.64)%

0.33&(.05P.80)%

0.37&(.04P1.15)%

0.55&(.05P1.70)%


0.29&(.10P.59)%

0.37&(.13P.74)%

0.41&(.11P.98)%

0.56&(.17P1.55)%

!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&wide& 0.44&

0.49&(.26P.84)%

0.55&(.26P.99)%

0.52&(.32P1.21)%

0.86&(.31P1.68)%

Albert&0.24&

0.25&(.06P.56)%

0.33&(.08P.60)%

0.26&(.14P.91)%

0.55&(.09P1.27)%

Kyoga&0.49&

0.57&(.25P1.00)%

0.60&(.27P1.13)%

0.67&(.41P1.31)%

0.99&(.40P1.86)%

Upper&Nile& 0.43&

0.43&(.21P.76)%

0.51&(.09P1.13)%

0.49&(.30P1.09)%

0.66&(.33P1.67)%


0.45&(.19P.85)%

0.50&(.20P1.02)%

0.50&(.22P1.46)%

0.92&(.26P1.96)%


0.66&(.23P1.16)%

0.73&(.18P1.29)%

0.80&(.15P1.78)%

0.17&(.17P2.39)%


0.62&(.29P1.10)%

0.69&(.33P1.26)%

0.74&(.31P1.56)%

1.09&(.35P2.26)%

!

FREQUENCY%OF%MONTHLY%PRECIPITATION%EXTREMES%

Shade&=&>%66%%of%models%project%an%increase%in%frequency%Shade%=%>%90%%of%models%project%an%increase%in%frequency%

Figure&29:&50%year%monthly%rainfall%event%in%mm/month%

Table&14:&Annual%frequency%of%monthly%precipitation%>300%mm%%

Table&12:&Annual%frequency%of%monthly%precipitation%>200%mm%%Monthly%Precipitation%Extremes:%The!frequency!that!monthly!precipitation!‘thresholds’!are!exceeded.###

• There#is#model#agreement#of!an!increase!in!extreme!monthly!precipitation!events.!#

• Table&13!shows!that!at!least!half!of!the!models!project!monthly!precipitation!of!250!mm/month!to!happen!at!least!three!times!more!often!than!it!has!historically.!!This!is!used!as!a!proxy!for!1997!precipitation!levels.!#

• Table&14&shows!that!at!least!half!of!the!models!project!the!highest!monthly!rainfall!on!record!to!occur!4!to!6!times!more!often!by!the!end!of!the!century.!!#

• Figure&29&shows!maps!of!the!50=year!event,!(i.e.!the!highest!monthly!rainfall!in!50!years)!It!is!shown!that!it!is!projected!to!increase!in!rainfall!throughout!this!century!under!both!emissions!scenarios!and!that!the!Lake!Victoria!region’s!50=year=event!received!nearly!twice!as!much!rainfall!as!other!region.#


! 20!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& 156.1&

171.5&(170P175)%

173.7&(169P177)%

176.3&(173P181)%

184.0&(181P192)%

Albert&148.0&

164.6&(163P168)%

166.3&(164P170)%

169.4&(166P173)%

176.6&(174P184)%

Kyoga&163.4&

179.4&(176P184)%

181.2&(177P186)%

184.1&(178P190)%

193.2&(188P201)%

Upper&Nile& 164.1&

178.0&(176P183)%

180.8&(171P185)%

184.0&(180P190)%

191.0&(190P200)%


161.7&(161P164)%

163.9&(162P167)%

166.1&(163P169)%

174.1&(171P180)%


163.7&(163P167)%

166.3&(164P169)%

168.2&(166P172)%

176.3&(173P182)%


162.0&(161P165)%

164.1&(162P167)%

166.2&(164P170)%

174.3&(171P181)%

!Shade&=&at%least%90%%of%models%project%an%increase%in%HWDI%

Figure&30:&The%median%model’s%projected%increase%in%PET%across%Uganda%(in%mm/month)%between%future%model%projections%and%historical%values.%%

POTENTIAL%EVAPOTRANSPIRATION%Potential%evapotranspiration%(PET):%%The!amount!of!evaporation!that!would!occur!if!there!were!a!sufficient!source!of!water!available!(mm/month).!(see!Annex!A!for!mathematical!details)%!

• All#models#project!an!increase!in!PET!over!the!next!century!!(see!Table&15)!and!an!increase!in!variation!of!PET!over!the!next!century!(See!Annex&B)!By!mid=century,!!Uganda!is!projected!to!see!an!increase!in!PET!between!7!and!21!mm/month.!!By!end=of=century,!PET!is!projected!to!increase!by!between!13!and!88!mm/month.!!

• Figure&30!shows!the!change!in!PET!varies#spatially!over!both!emissions!scenarios!for!mid=century!and!end=of=century.!!The!northern!part!of!Uganda!is!projected!to!see!a!slightly!higher!increase!in!PET.!!

• Figure&31&are!boxplots!of!modeled!monthly!projections!of!PET,!plotted!on!top!of!historical!monthly!precipitation.!!This!plot!shows!that!PET!for!every!month!exceeds!historical!median!monthly!precipitation!levels.!!For!April,!May,!September!and!October,!the!75th!percentile!of!historical!monthly!precipitation!intersects!the!monthly!PET!boxes.!!With!uncertainty!about!the!direction!of!change!of!monthly!precipitation!and!certainty!about!the!increase!PET,!we!expect!there!to!be!a!decrease!in!available!water!due!to!increased!evaporation!and!insufficient!increases!rainfall!to!replenish!the!increased!loss!of!water.!&

Figure&31:&Boxplots%of%model%PET%projections%by%month.%%The%black%solid%line%is%historical%monthly%mean%precipitation.%%The%dotted%blue%lines%are%the%25th%and%75th%percentile%of%historical%monthly%mean%projections.%%%


! 21!

Figure&33a:&Annual%frequency%of%low%extremes%(WASP%<%P1)%

Figure&33b:&Annual%frequency%of%high%extremes%(WASP%>1.5)%

Weighted%Anomaly%Standardized%Precipitation%(WASP):%%An!annual!index!of!monthly!rainfall!deviation!from!monthly!climatology.!!Positive!values!indicate!above!normal!rainfall!levels,!negative!values!indicate!below!normal!rainfall!levels!(see!Annex!A!for!more!details).!!

• Most#models#project!an!increase!in!high!WASP!values!(See!Figure&32),!indicating!a!projected!increase!in!frequency!of!above!normal!monthly!precipitation!extremes.%

• Figure&33a!shows!that!most!models!project!a!decrease!in!frequency!of!low!WASP!values!throughout!the!century!with!higher!emissions!scenarios!leading!to!lower!WASP!values.!This!figure!also!shows!that!WASP#values#vary#spatially,!as!do!changes!in!WASP!values.%

• Figure&33b&shows!that&most!models!project!an!increase!in!the!frequency!of!high!WASP!values!throughout!the!century!with!higher!emissions!scenarios!leading!to!higher!frequencies!of!high!WASP!values.%

• Figure&33a&and!33b&show!that!models!are!projecting!changes!in!low!WASP!values!to!correspond!regionally!with!changes!in!high!WASP!values!indicating!that!the!range!of!rainfall!values!is!shifting!towards!higher!extremes!(i.e.!higher!WASP!values).%

Figure&32:&Boxplots%of%projected%WASP%values.%%Each%box%corresponds%to%one%model’s%annual%WASP%values%for%the%given%time%period%and%emissions%scenario.%%The%black%dotted%lines%are%the%interquartile%range%for%historical%annual%WASP%values.%%%

WEIGHTED%ANOMALY%STANDARDIZED%PRECIPITATION%%

! 22!

&&Region&

1950:2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&wide& 53.1%

59.9&(42.2P74.9)%

64.5&(45.6P82.6)%

68.1&(39.9P122)%

76.6&(44.9P157)%

Albert&55.4%

61.4&(48.5P78.4)%

64.3&(46.6P85.8)%

67.6&(43.6P124)%

76.6&(45P144)%

Kyoga&43.9%

54.1&(36.9P70.0)%

55.4&(38.5P79.6)%

61.9&(37.2P123)%

76.7&(37.4P170)%

Upper&Nile& 24.8%

27.0&(21.4P38.9)%

31.0&(21.6P45.6)%

32.4&(19.0P66.2)%

39.3&(23.8P82.8)%

Victoria&Land& 79.2%

87.3&(64.9P111)%

90.8&(66.8P116)%

96.8&(61.4P163)%

106.5&(64.2P198)%

Victoria&Uganda& 86.8%

100.3&(59.9P122)%

100.3&(71.6P132)%

109.1&(53.9P177)%

116.9&(58.2P233)%

Victoria&Basin& 96.1%

111.8&(79.6P132)%

109.8&(83.1P139)%

120.2&(78.7P183)%

129.0&(79.2P228)%

!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&wide& 50%% 44%& 47%& 49%& 57%&Albert& 43%% 38%& 36%& 42%& 47%&Kyoga& 61%% 49%& 51%& 58%& 70%&Upper&Nile& 62%% 61%& 65%& 68%& 72%&Victoria&Land& 35%% 27%& 27%& 33%& 39%&Victoria&Uganda& 35%% 28%& 28%& 33%& 38%&Victoria&Basin& 33%% 28%& 29%& 33%& 38%&!

DECEMBERPFEBRUARY%PRECIPITATION%

Figure&35:&DJF%projected%change%in%precipitation%where%66%%of%models%agree%on%direction%of%change%

Figure&34:&DJF!projected!precipitation!distribution!

December?February%(DJF)%Precipitation:!precipitation!reported!as!a!monthly!average!over!three!months.!!!

• Most#models#project!an!increase!in!DJF!precipitation!across!Uganda!for!both!emissions!scenarios!and!time!periods!(Table&16).!!

• The!median!model!projects!the!distribution!of!DJF!precipitation!to!shirt!up!(see!Figure&34)!with!some!models!projecting!significant!increases!in!DJF!precipitation!towards!the!end!of!the!century!

• Figure&35!shows!that!the!change!in!DJF!precipitation!is!projected!to!vary#spatially!with!higher!increases!in!precipitation!in!the!southwestern!part!of!the!country.!!!This!figure!also!shows!that!the!greatest!uncertainty!in!DJF!precipitation!among!models!is!in!the!western!border!of!the!country.!!!

• !Table&17&shows!that!models!project!a!decrease!in!the!coefficient!of!variation!for!DJF!precipitation!for!mid!century!under!the!moderate!emissions!scenario!in!all!WMZs!except!the!Upper!Nile!meaning!they!are!projecting!precipitation!to!be!less!variable!for!these!WMZs.!!!

Table&16:&DJF%Average%monthly%precipitation%%

Table&17:&Coefficient%of%Variation%of%DJF%monthly%precipitation%%%

Shade&=&>%66%%of%models%project%an%increase%in%DJF%precipitation%Shade%=%>%90%%of%models%project%an%increase%in%DJF%precipitation%

Shade&=&>%66%%of%models%project%an%increase%in%CV%of%DJF%precipitation%Shade%=%>%66%%of%models%project%an%decrease%in%CV%of%DJF%precipitation%

! 23!

&&Region&

1950:2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&wide& 136.8%

133.2&(118P150)%

129.4&(110P176)%

132.2&(116P171)%

138.5&(99.4P195)%

Albert&126.3%

121.3&(109P134)%

121.9&(104P152)%

120.4&(103P155)%

126.6&(106P176)%

Kyoga&136.1%

138.6&(118P157)%

129.2&(104P183)%

138.2&(119P172)%

135.7&(99.9P197)%

Upper&Nile& 112.6%

111.8&(93.5P130)%

109.9&(77.8P145)%

112.8&(91.1P143)%

113.4&(97.6P158)%

Victoria&Land& 154.8%

146.3&(130P165)%

147.0&(134P203)%

146.4&(130P204)%

150.4&(108P236)%

Victoria&Uganda& 170.4%

163.4&(134P181)%

160.9&(128P221)%

163.7&(128P217)%

168.3&(92.6P248)%


151.9&(131P174)%

146.0&(129P216)%

148.8&(135P211)%

157.5&(108P251)%

!

&&Region&

1950:

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda:wide& 22%% 22%% 24%% 23%& 26%&Albert& 20%% 21%% 23%% 21%& 24%&Kyoga& 25%% 26%% 26%% 27%& 29%&Upper&Nile& 24%% 21%% 23%% 22%& 25%&Victoria:Land& 19%% 20%% 21%% 20%& 24%&Victoria:in&Uganda& 20%% 20%% 22%% 22%& 25%&Victoria&Basin& 24%% 26%% 30%% 28%& 31%&!

MARCHPMAY%PRECIPITATION%March?May%(MAM)%Precipitation:!precipitation!reported!as!a!monthly!average!over!three!months.!!!

• Most#models#project!a!decrease!in!precipitation!during!MAM!for!the!Albert!and!Victoria!WMZ!!(Table&18)!for!all!scenarios!except!the!high!emissions!scenario!at!the!end!of!the!century.!

• The!median!model!is!projecting!the!distribution!in!monthly!MAM!precipitation!to!shift!downwards!(Figure&36).!!Some!models,!however,!project!the!distribution!of!precipitation!to!shift!up!considerably.!!!

• There!is!uncertainty!among!models!regarding!the!direction!of!change!for!MAM!for!most!of!Uganda,!however,!Figure&37!shows!that!most!models!agree!there!will!be!a!decrease!in!precipitation!in!the!southwestern!part!of!the!country.!!!

• By!the!end!of!the!century!under!the!high!emissions!scenario,!most!models!project!an!increase!in!variation!in!MAM!monthly!precipitation!for!all!regions!except!Upper!Nile!where!there!is!disagreement!among!models.!!!

Shade&=&>%66%%of%models%project%an%decrease%in%MAM%precipitation% Shade&=&>%66%%of%models%project%an%increase%in%CV%of%MAM%precipitation%

Table&19:&Coefficient%of%Variation%of%MAM%monthly%precipitation%%%

Table&18:&MAM%Average%monthly%precipitation%%

Figure&36:&MAM%projected%precipitation%distribution%

Figure&37:&DJF%projected%change%in%precipitation%where%66%%of%models%agree%on%direction%of%change%

! 24!

&&Region&

1950:2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda:wide& 99.9%

98.2&(77.9P128)%

94.3&(73.6P126)%

98.9&(74.0P131)%

103.0&(69.8P142)%

Albert&83.9%

81.2&(66.7P114)%

82.1&(62.5P112)%

87.2&(65.4P114)%

86.8&(59.6P122)%

Kyoga&109.5%

112.7&(77.8P143)%

105.3&(76.3P140)%

111.4&(76.4P140)%

117.9&(64.8P189)%

Upper&Nile& 139.3%

134.0&(106P168)%

130.0&(98.1P174)%

135.5&(109P172)%

139.4&(99.1P182)%

Victoria:Land& 60.1%

58.5&(45.2P83.7)%

57.4&(41.0P79.9)%

61.4&(39.2P89.0)%

64.4&(35.4P103)%

Victoria:Uganda& 67.6%

67.1&(39.7P87.7)%

64.2&(42.6P86.4)%

68.8&(33.7P98.9)%

71.9&(40.7P115)%


50.2&(33.7P65.7)%

50.3&(36.1P69.3)%

52.3&(29.9P76.3)%

49.2&(33.1P99.4)%

!

&&Region&

1950:2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda:wide& 31%% 42%% 43%% 46%% 48%%Albert& 29%% 39%% 39%% 49%% 42%%Kyoga& 32%% 44%% 41%% 46%% 47%%Upper&Nile& 29%% 37%% 35%% 36%% 36%%Victoria:Land& 34%% 49%% 52%% 58%% 57%%Victoria:Uganda& 32%% 48%% 51%% 54%% 55%%Victoria&Basin& 54%% 74%% 77%% 80%% 88%%!

JUNEPAUGUST%PRECIPITATION%June?August%(JJA)%Precipitation:!precipitation!reported!as!a!monthly!average!over!three!months.!!!

• There!is!disagreement!among!models!with!regards!to!the!direction!of!change!on!JJA!precipitation!(Table&20).!!

• Most!models!agree!that!there!will!be!!an!increase!in!the!coefficient!of!variation!for!JJA!monthly!precipitation!by!at!least!10%!(Table&21).!!!

• Figure&38&shows!that!the!median!model!projects!an!increase!in!frequency!of!low!precipitation!months!and!an!increase!in!frequency!of!high!precipitation!months!for!JJA!with!a!decrease!in!frequency!of!mid=range!precipitation!months.!!(an!increase!in!variability!in!precipitation)!

• Figure&39&shows!that!there!is!uncertainty!among!models!with!regard!to!the!direction!of!change!on!a!spatial!level.!!By!mid=century!most!models!agree!there!will!be!a!slight!decrease!in!precipitation!in!the!southern!Lake!Victoria!region.!!End!of!century!change!is!even!less!certain.!!

Table&21:&Coefficient%of%Variation%of%JJA%monthly%precipitation%%%

Shade&=&>%66%%of%models%project%an%increase%in%CV%of%JJA%precipitation%Shade%=%>%90%%of%models%project%an%increase%in%CV%of%JJA%precipitation%Shade&=&>%66%%of%models%project%an%decrease%in%JJA%precipitation%

Table&20:&JJA%Average%monthly%precipitation%%

Figure&39:&JJA%projected%change%in%precipitation%where%66%%of%models%agree%on%direction%of%change%

Figure&38:&JJA%projected%precipitation%distribution%

! 25!

&&Region&

1950:2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda:wide& 125.9%

126.9&(105P157)%

125.2&(94.8P155)%

132.1&(98.3P163)%

147.5&(96.5P176)%

Albert&143.5%

144.2&(117P175)%

142.6&(110P179)%

148.8&(122P180)%

159.5&(116P193)%

Kyoga&111.9%

123.7&(92.2P144)%

118.9&(77.3P139)%

124.1&(81.0P152)%

141.0&(84.2P175)%

Upper&Nile& 120.7%

121.2&(99.8P155)%

120.7&(88.5P155)%

122.5&(89.1P153)%

143.8&(89.7P180)%


128.8&(111P166)%

131.3&(95.0P167)%

130.3&(106P172)%

146.1&(97.8P189)%


127.7&(109P172)%

131.2&(91.6P162)%

138.6&(101P182)%

148.3&(95.5P193)%


103.9&(86.7P140)%

107.3&(71.7P133)%

115.3&(83.8P146)%

120.9&(78.2P163)%

!

&&Region&

1950:2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda:wide& 32%% 33%% 36%% %%%%%37%% 39%%Albert& 24%% 25%% 26%% 25%% 28%%Kyoga& 41%% 46%% 48%% 48%% 51%%Upper&Nile& 31%% 36%% 38%% 39%% 40%%Victoria:Land& 28%% 29%% 29%% 29%% 34%%Victoria:Uganda& 32%% 32%% 33%% 34%% 35%%Victoria&Basin& 38%% 42%% 43%% 45%% 47%%!

SEPTEMBERPNOVEMBER%PRECIPITATION%September?November%(SON)%Precipitation:!precipitation!reported!as!a!monthly!average!over!three!months.!!!

• There!is!uncertainty!among!models!with!regards!to!the!direction!of!change!for!SON!precipitation!(Table&22).!

• There!are!some!regions!and!time!periods!where!most!models!project!an!increase!in!CV!(Table&23).!

• Figure&40!shows!that!the!median!model!projects!the!distribution!of!SON!precipitation!to!flatten!out!with!a!higher!frequency!of!high!precipitation!months.!!The!error!bars!show,!however,!that!there!are!some!model!that!predict!a!greater!frequency!of!mid=range!precipitation!months.!

• Figure&41&shows!that!the!direction!of!change!is!uncertain!among!models!on!a!spatial!level.!!However,!models!agree!that!by!mid!century!there!will!be!a!small!increase!in!precipitation!in!the!central!eastern!part!of!the!country!and!that!by!the!end!of!the!century!there!is!a!greater!region!of!certainty!among!models!of!an!increase!in!precipitation.!

Table&23:&Coefficient%of%Variation%of%SON%monthly%precipitation%%%

Shade&=&>%66%%of%models%project%an%increase%in%CV%of%SON%precipitation%Shade&=&>%66%%of%models%project%an%increase%in%SON%precipitation%

Table&22:&SON%Average%monthly%precipitation%%

Figure&41:&SON%projected%change%in%precipitation%where%66%%of%models%agree%on%direction%of%change%

Figure&40:&SON%projected%precipitation%distribution%

! 26!NUMBER%OF%WET%DAYS%A%YEAR%

&&Region&

1950:2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda:wide& 117.3%

121.7&(109P132)%

114.8&(111P130)%

129.6&(110P136)%

136.2&(113P143)%

Albert&133.9%

134.5&(124P149)%

129.3&(126P144)%

142.8&(122P150)%

151.6&(126P162)%

Kyoga&102.2%

113.5&(93.5P118)%

103.3&(95.5P118)%

115.6&(95.9P135)%

123.7&(96.2P132)%

Upper&Nile& 89.1%

89.4&(82.3P108)%

87.0&(71.8P102)%

99.0&(82.1P108)%

93.8&(89.5P118)%


147.6&(140P164)%

147.8&(139P162)%

159.7&(137P164)%

168.6&(137P181)%


149.1&(136P161)%

142.4&(135P157)%

154.9&(135P166)%

163.3&(137P176)%


135.9&(129P147)%

132.1&(122P148)%

138.5&(128P159)%

146.7&(127P164)%

!Shade&=&>%66%%of%models%project%an%increase%in%number%of%wet%days%a%year%Shade%=%>%90%%of%models%project%an%increase%in%number%of%wet%days%a%year%

Table&24:&Number%of%Wet%Days%a%Year%%

Number%of%Wet%Days%a%Year:%The!number!of!days!a!year!where!precipitation!is!at!least!1mm/day.!!!!

• Most#models!agree!that!by!the!end!of!century!there!will!be!an!increase!in!the!number!of!wet!days!in!Uganda!(Table&24).!

• Figure&42&shows!that!the!distribution!of!the!number!of!wet!days!a!year!is!flattening!out.!!The!median!model!projects!there!to!be!years!with!80!and!90!days!of!precipitation!where!historically!there!have!been!no!fewer!than!100!wet!days!in!Uganda.!!The!median!model!also!projects!there!to!be!an!increase!in!frequency!of!years!where!there!are!greater!than!150!wet!days!in!Uganda.!!This!figure!suggests!that!models!are!predicting!greater!variability!in!the!number!of!wet!days.!!!

• Figure&43!shows!that!there!is!some!spatial!agreement!among!models!with!regards!to!the!change!in!the!number!of!wet!days.!!By!the!end!of!the!century!most!models!agree!there!will!be!little!change!in!the!lake!Victoria!basin!and!an!increase!in!wet!days!throughout!the!central!part!of!the!country.!!!

Figure&42:%Median%model%projected%distribution%in%number%of%wet%days%a%year%

Figure&43:%Median%model%projected%change%in%number%of%wet%days%a%year%where%at%least%66%%of%models%agree%on%either%direction%of%change%or%on%there%being%no%change.%%Grey%means%less%than%66%%of%models%agree.%

! 27!

&&Region&

1950&:&

2000&


High&emissions&

Moderate&emissions&

High&emissions&

Uganda&& :0.35&

:0.40&(P0.43,%P0.25)%

:0.40&(P0.47,%P0.23)%

:0.38&&&(P0.42,%P0.21)%

:0.33&(P0.40,%P0.09)%

Albert&:0.33&

:0.39&(P0.46,%P0.26)%

:0.39&(P0.47,%P0.27)%

:0.39&(P0.46,%P0.23)%

:0.32&(P0.42,%P0.15)%

Kyoga&:0.40&

:0.41&(P0.47,%P0.27)%

:0.44&(P0.49,%P0.29)%

:0.35&(P0.47,%P0.27)%

:0.37&(P0.41,%P0.13)%

Upper&Nile& :0.41&

:0.44&(P0.54,%P0.25)%

:0.43&(P0.61,P0.23)%

:0.45&(P0.53,%P0.19)%

:0.33&(P0.52,%P0.06)%

Victoria&Land& :0.29&

:0.35&(P0.40,%P0.23)%

:0.36&(P0.38,%P0.20)%

:0.35&(P0.38,%P0.17)%

:0.29&(P0.39,%P0.08)%

Victoria&Uganda& :0.26&

:0.32&(P0.34,%P0.17)%

:0.31&(P0.33,%P0.16)%

:0.27&(P0.33,%P0.14)%

:0.26&(P0.33,%P0.04)%

Victoria&Basin& :0.32&

:0.36&(P0.42,%P0.26)%

:0.36&(P0.40,%P0.21)%

:0.33&(P0.37,%P0.19)%

:0.30&(P0.37,%P0.09)%

!

CLIMATE%MOISTURE%INDEX%

Figure&44:&Boxplots%of%CMI.%%Each%model’s%projections%are%separated%into%individual%boxes.%%The%horizontal%lines%are%the%historical%median%with%25th%and%75th%percentile%on%either%side.%

Figure&45:&CMI%for%historical%and%median%model%future%projections.%%%

Figure&46:&Maps%where%>66%%of%models%agree%on%the%direction%of%change%of%CMI.%%Grey%means%that%<66%%of%agree%on%direction%of%change.%

Climate%Moisture%Index%(CMI):!A!measure!of!aridity!derived!from!PET!and!precipitation!and!on!a!scale!of!=1!to!1!where!negative!values!signal!dryness!and!positive!values!signal!high!moisture!levels!(see!Annex!A!for!details).!!

• Most#models!project!a!decrease!in!CMI!values!by!mid=century!on!a!country=wide!level!though!there!is!uncertainty!about!the!direction!of!change!for!the!Kyoga!and!Upper!Nile!WMZ!(Table&25).!

• Figure&44&shows!that!a!few!models!project!an!increase!in!range!of!CMI!values!though!most!project!a!range!of!CMI!values!similar!to!the!historical!range.!!

• Figure&45!shows!that!CMI!varies!spatially!with!the!part!of!the!northeastern!part!of!the!country!and!southeastern!part!of!the!Victoria!basin!having!the!lowest!CMI!values.!!This!figure!also!shows!that!the!median!model!projects!a!slight!decrease!in!CMI!in!some!parts!of!the!country!by!mid=century!(a!result!of!warming)!but!an!increase!towards!the!end!of!the!century!(a!result!of!increased!precipitation).!!&

• Figure&46&shows!that!spatially,!models!disagree!on!the!direction!of!change!though!there!is!some!spatial!agreement!among!models!on!a!low!level!or!no!change!in!CMI.&

Shade&=&where!>66%!of!models!agree!on!a!decrease!in!CMI!

=1! ! ! ! !!!!0! ! ! ! !!!!!!!!!!1!

Table&25:&Climate%Moisture%Index%%

!

!

28!

(CONCLUSIONS! !!Uganda!already!has!a!highly!variable!climate,!over!both!space!and!time.!!This!analysis!indicates!that!temporal!and!spatial!variability!for!a!number!of!key!indicators!will!become!even!more!pronounced!in!the!future.!!Broadly,!the!findings!of!the!analysis!are!the!following:!!• Models!agree!that!the!climate!in!Uganda!will!warm!over!the!next!century.!!There!is!projected!to!be!an!

increase!in!mean,!minimum,!and!maximum!daily!temperature.!Models!also!agree!that!a!higher!emissions!scenario!will!result!in!a!warmer!climate!and!that!temperatures!stop!increasing!towards!the!end!of!the!century!under!a!moderate!emissions!scenario.!!

!• With!warmer!temperatures,!Uganda!is!projected!to!face!higher!levels!of!potential!evapotranspiration,!

leading!to!a!more!arid!climate!throughout!the!country.!!Heat!waves!are!projected!to!increase!in!frequency!and!to!last!longer!as!the!century!unfolds.!!Projections!indicate!that!there!will!also!be!more!days!where!temperature!exceeds!30°C!throughout!the!country.!!!

!• There!is!less!agreement!amongst!models!on!the!direction!of!change!in!annual!precipitation.!!However,!

all!models!project!precipitation!to!become!more!variable,!with!more!precipitation!extremes!and!longer!dry!periods.!!The!most!extreme!precipitation!events!witnessed!over!the!last!50!years!–!namely,!the!highest!monthly!precipitation!on!record!(in!1962)!and!the!second!highest!(in!1997),!which!led!to!extensive!flooding!in!the!country!–!are!likely!to!occur!more!often!and!even!be!exceeded!through!the!21st!century.!!Seasonally,!the!December:February!dry!season!is!projected!to!become!wetter!and!March:May!wet!season!is!projected!to!become!drier.!!!

!• Climate!change!will!vary!spatially.!!Temperatures!are!projected!to!increase!more!in!the!southern!part!

of!the!country,!where!the!projections!also!indicate!a!greater!increase!in!the!number!of!consecutive!dry!days.!!There!is!a!higher!level!of!model!agreement!that!DJF!precipitation!will!increase!most!around!Lake!Victoria!region!and!that!MAM!precipitation!will!decrease!in!the!southwestern!corner!of!Uganda.!!

!This!report!is!consistent!with!some!of!the!findings!of!previous!work!on!climate!change!in!Uganda.!!For!example,!projections!of!temperature!changes!in!this!analysis!are!similar!to!those!found!by!Baastel!(2014),!USAID!(2013)!and!DEWPoint!(2012).!!!Projections!of!precipitation!are!consistent!with!the!USAID!(2013)!report,!which!found!an!increase!in!precipitation!during!the!December!to!February!season.!!This!analysis!also!confirms!previous!findings!that!extremes!are!projected!to!increase!throughout!Uganda,!while!considering!a!wider!range!of!indicators!to!support!this!conclusion.!!However,!there!are!some!differences!between!the!findings!of!this!analysis!and!earlier!efforts.!!For!example,!Baastel!(2014)!finds!that!models!project!a!decrease!in!precipitation.!!The!conclusion!is!not!supported!when!a!larger!set!of!GCMs!is!analyzed!as!is!recognized!best!practice!and!the!approach!taken!here.!This!analysis!also!builds!on!and!updates!earlier!work!on!climate!change!in!Uganda!by!considering!a!wider!range!of!climate!indicators!using!the!latest!IPCC!emissions!scenarios.!!While!this!analysis!is!likely!the!most!comprehensive!and!rigorous!for!Uganda!to!date,!there!remain!areas!for!refinement!as!is!the!case!for!all!such!efforts.!First,!a!historical!dataset!that!ends!in!2008!is!used.!!Updating!the!dataset!to!include!the!most!recent!years!of!observations!would!improve!the!historical!analysis.!!Second,!the!analysis!would!benefit!from!a!closer!examination!and!inclusion!of!available!!station!data!for!Uganda!(see!Annex!E!for!a!statistical!comparison!with!available!station!data).!!Third,!this!work!could!be!extended!by!applying!climate!output!from!the!GCMs!to!a!rainfall:runoff!model!in!order!to!better!understand!how!projected!changes!affect!various!hydrologic!indicators,!such!as!floods!and!river!discharge.!!Finally,!while!this!analysis!is!intended!to!provide!insights!into!current!and!future!climatic!

!

!

29!

conditions!in!Uganda,!it!is!not!a!climate!impact!or!risk!assessment.!!More!work!would!be!required!to!translate!the!range!of!climate!outcomes!presented!here!into!climate!impacts!and!risks!at!various!levels.!!This!work!contributes!to!a!growing!knowledge!base!for!understanding!current!and!future!climate!in!Uganda.!!!Notwithstanding!the!rigor!of!the!analysis,!there!is!substantial!uncertainty!about!the!state!of!future!climate.!!Although!the!greatest!range!of!possible!future!climate!outcomes!for!Uganda!to!date!is!evaluated,!this!analysis!does!not!capture!all!possible!futures!because!it!is!based!on!a!sub:set!of!GCM:emissions!scenarios!as!well!as!a!sub:set!of!all!possible!GCM!projections.!!The!results,!however,!clearly!show!that!Uganda!is!already!contending!with!a!high!degree!of!climate!variability!and!that!this!is!expected!to!increase!in!the!future.!!The!implication!is!that!development!planning!must!carefully!consider!and!fully!integrate!the!variability!of!today’s!climate,!as!it!presents!challenges!that!have!yet!to!be!adequately!addressed.!!Adapting!to!current!climatic!conditions!will!prepare!Uganda!to!be!able!to!manage!potentially!more!complex!–!and!uncertain!–!climate:related!risks!in!the!future.!(((((((((((((((((((((((((

!

!

30!

(REFERENCES(!IPCC!2014!WGII!AR5!!Summary!for!Policy!Makers!In:!Chapter*22.*Africa,![Niang,!I.,!Ruppel,!O.C.,!Abdrano,!M.;!Essel,!A.,!Lennard,!C.,!Padgham,!J.,!Urqhart,!P.]!Cambridge!University!Press,!Cambridge,!United!Kingdom!and!New!York,!NY,!USA!!Baastel!May!2014.!Economic*Assessment*of*the*Impacts*of*Climate*Change*in*Uganda:*Regional9scale*Climate*Change*Projections*of*Annual,*Seasonal*and*Monthly*Near9Surface*Temperatures*and*Rainfall*in*Uganda,*[Rautenbach,!H.,!Botaï,!J.,!Wasswa,!F.,!Shongwe,!M.,!Beucher,!O.]!!!Tetra!Tech!August!2013.!Uganda*Climate*Change*Vulnerability*Assessment*Report,![Caffrey,!P.,!Finan,!T.,!Trzaska,!S.,!Miller,!D.,!Laker:Ojok,!R.,!Huston,!S.]!United!States!Agency!for!International!Development!by!Tetra!Tech!ARD!!LTS!2008.!Climate*Change*in*Uganda:*Understanding*the*Implications*and*Appraising*the*Response.*![Hepworth,!N.,!Goulden,!M.]!LTS!International,!Edinburgh!!LTS!2010.!!Climate*Change*Vulnerability*and*Adaptation*Preparedness*in*Uganda*[Hepworth,!N.]!LTS!Africa,!Heinrich!Böll!Foundation,!Nairobi,!Kenya!!UNDP!2004.!Policy*Brief:*Climate*Change,*The*environment*and*Human*Welfare:*Lessons*Learned*from*the*Lake*Kyoga*Catchment*Area,*UNDP:UNEP!Poverty:Environment!Initiative!!DEWPoint!(2012)!Support!to!the!Strategic!Programme!Review!for!Climate!Change!in!Uganda:Understanding!the!implications!and!appraising!the!response.!Update!to!2008!LTS!Scoping!Study![Steynor,A.,!Jack,!C.!and!Smith,!C.],!DFID!Resource!Centre!for!Environment,!Water!and!Sanitation,!51pp.!(NPA!(2014)!Integration*of*Climate*Change*Policy*Concerns*into*the*Second*National*Development*Plan!(NDP!II),![Dr.!Revocatus!Twinomuhangi,!REMODE!Consults!Limited,!Makerere!University]!National!Planning!Authority,!The!Republic!of!Uganda,!!!Rogelj!(2012).!Global*warming*under*old*and*new*scenarios*using*IPCC*climate*sensitivity*range*estimates,*Rogelj,!J.,!Meinshausne,!M.,!Knutti,!R.,!Nature!Climate!Change,!Vol!2,!April!2012!p!248:253

!

!

31!

ANNEX(A((Equations!for!climate!indicators!are!derived!from!precipitation!or!temperature!on!a!monthly!or!daily!time!step.!!For!daily!data,!maximum!daily!temperature!and!minimum!daily!temperature!are!provided!by!model!projections!and!historical!data.!!Formal!calculations!are!provided!here!for!three!indicators!that!are!more!complex!to!derive.!!!*Potential(Evapotranspiration(Potential!evaporation!is!calculated!for!each!day!and!takes!into!account!the!latitude!and!time!of!year!as!well!as!temperature,!temperature!range!and!and!precipitation!and!uses!the!‘modified!Hargreaves’!approach:!!

(!"# = 0.0013 − 0.408!" ∙ (!!"# − 17.8) ∙ (!" − 0.0123!)!.!",(

(where,!RA!! =!incoming!solar!radiation!(determined!by!time!of!year!and!latitude),!!!"#=!mean!daily!temperature!TD!! =!temperature!range!for!that!day,!(i.e.!Tmax!–!Tmin)!P!!! =!!precipitation!in!mm.!!!Weighted(Anomaly(Standardized(Precipitation(Index(WASP!is!calculated!for!each!year!in!the!sample!and!is!summed!over!the!twelve!months!in!that!calendar!year.!

!"#$! =1!!

(!! − !!!!

∙ !!!!)

!"

!!!!

where,!!!!!=!the!standard!deviation!of!monthly!precipitation!for!the!m!month!of!the!year!over!all!years.!!!!!=!monthly!precipitation!for!month!m*in!year!y!!!!=mean!monthly!precipitation!for!month!m*!over!all!years!!!!=!average!annual!precipitation!over!all!years!!!!!=!standard!deviation!of!the!twelve!values!being!summed.!!(Climate(Moisture(Index.((CMI!is!calculated!for!each!year!and!using!annual!values!for!PET!and!P*(precipitation).!!!If!P*>*PET!then:!! ! CMI*=*19PET/P!If!P*<*PET!then:!! ! CMI*=*P/PET*9*1!(

!

!

32!

ANNEX(B(The!following!tables!report!of!the!coefficient!of!variation!for!Monthly!precipitation!(Table"27)!and!potential!evapotranspiration!(Table"28).!!!!Table"27!shows!that!averaged!over!the!country,!most!models!agree!that!precipitation!will!become!more!variable!towards!the!middle!of!the!century.!!!Spatially,!the!Victoria!and!Albert!WMZs!are!projected!to!become!more!variable!in!terms!of!monthly!precipitation!meaning!that!monthly!precipitation!will!deviate!from!the!mean!by!more!or!more!often!than!historical!deviation.!!!!Table"28!shows!that!nearly!all!models!anticipate!the!variation!to!increase!for!Potential!Evapotranspiration!substantially!under!both!time!periods!and!both!RCP!scenarios.!!!!Table&27:&Coefficient(of(variation(for(Monthly(Precipitation(&&Region&

&&

Historical&

2030T2050& 2080T2100&Moderate&emissions&

High&emissions&

Moderate&emissions&

High&emissions&

UgandaTwide& 67%& 69%& 72%& 70%& 69%&Albert& 59%& 61%& 62%& 61%& 61%&Kyoga& 73%& 76%& 77%& 75%& 77%&Upper&Nile& 75%& 75%& 78%& 75%& 73%&VictoriaTLand& 60%& 64%& 65%& 65%& 66%&VictoriaTin&Uganda& 61%& 64%& 67%& 65%& 66%&Victoria&Basin& 73%& 76%& 78%& 79%& 80%&

Blue(indicates(where(at(least(66%(of(models(agree(on(an(increase(in(CV((Pink(indicates(where(at(least(66%(of(models(agree(on(a(decrease(in(CV(

((Table&28:&Coefficient(of(variation(for(PET(&&Region&

&&

Historical&

2030T2050& 2080T2100&Moderate&emissions&

High&emissions&

Moderate&emissions&

High&emissions&

UgandaTwide& 3%& 11%& 12%& 12%& 15%&Albert& 3%& 9%& 9%& 10%& 12%&Kyoga& 4%& 12%& 13%& 13%& 16%&Upper&Nile& 5%& 16%& 16%& 17%& 19%&VictoriaTLand& 2%& 7%& 8%& 8%& 10%&VictoriaTin&Uganda& 2%& 8%& 8%& 8%& 11%&Victoria&Basin& 2%& 9%& 9%& 9%& 11%&

Blue(indicates(where(at(least(90%(of(models(agree(on(an(increase(in(CV(

(""""

!

!

33!

"ANNEX(C"The dataset used in this profile for projecting climate change impacts for Uganda is a set of Bias-Corrected and Spatially Disaggregated (BCSD) climate projections based on GCM runs from the Coupled Model Intercomparison Project phase 5 (CMIP5) IPCC archives. In this Annex, the steps are described in generating the BCSD projections.

The bias corrected climate projections are produced for the African continent as part of the ongoing Vulnerability of Africa’s Infrastructure to Climate Change study. From this ongoing work, there are a total of 121 projections available with a monthly time step and 95 available with a daily time step. Of these, 65 monthly projections and 39 daily projections are GCMs from the CMIP5 ensemble. Of these projections, 23 monthly and 9 daily GCMs run under RCP 4.5, and 20 monthly and 8 daily GCMs run under RCP 8.5 were bias corrected to historical data. The GCM time periods vary, but all were available between 1900 and 2100.

METHODOLOGY

The methodology used to generate the BCSD projections closely follows that outlined by Maurer and others as part of a project jointly funded by multiple U.S. agencies to bias correct and downscale climate and hydrology projections based on the CMIP3 archive. Results of this project and methods employed are documented on a website hosted by the Lawrence Livermore National Laboratory (LLNL).4

The approach used for this profile involves four steps: (1) select the baseline dataset and GCM-emissions combinations, (2) resolve GCMs and baseline to a common spatial resolution, (3) spatially disaggregate those outputs, and (4) bias correct the GCM outputs. Note that LLNL bias corrects the data before doing spatial disaggregation, whereas we start with spatial disaggregation and then bias correct.5

STEP 1: SELECT OBSERVED BASEL INE DATASET AND GCM RUNS

The first step is to select the observed baseline and set of GCM-emissions combinations from available IPCC archives. The observed baseline dataset employed in this analysis was generated by the Princeton Land Surface Hydrology Research Group (henceforth, Princeton dataset), and is available globally from 1948 to 2008 at a monthly or daily time step at 0.5 x 0.5 degree for average temperature (tmean) and precipitation, and at 1 x 1 degree for maximum and minimum daily temperature (tmax and tmin).6 To produce baseline tmax and tmin at a 0.5 x 0.5 degree resolution, it is assumed that the information contained in each 1 x 1 degree Princeton grid cell can be transferred to the underlying four 0.5 x 0.5 degree grid cells. For each month and 1 x 1 degree grid cell, first the difference between tmean and tmin, and between tmax and tmean is calculated and then applied uniformly to the tmean value in each of the four corresponding 0.5 x 0.5 degree grids.

STEP 2: RESOLVE BASEL INE AND GCMS TO COMMON SPATIAL RESOLUTION

Spatial resolutions of the GCMs range from approximately 1 x 1 degree to 4 x 5 degrees. Following the first step in the BCSD process outlined by Maurer and others, each of the GCMs and the Princeton baseline are !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!4!See!“Bias!Corrected!and!Downscaled!WCRP!CMIP3!Climate!and!Hydrology!Projections”!at!http://gdo:dcp.ucllnl.org/downscaled_cmip_projections/dcpInterface.html#About!(accessed!on!April!28,!2013)!for!more!information!and!sources!documenting!the!BCSD!approach.!5!According!to!Dr.!Maurer,!although!spatially!disaggregating!first!has!analytical!advantages,!the!LLNL!team!elected!to!start!with!bias!correction!because!of!restrictions!on!computing!power!(i.e.,!starting!with!disaggregation!requires!bias!correcting!16!times!as!many!grid!cells).!Based!on!personal!communication!with!Edward!Maurer!(Associate!Professor!at!Santa!Clara!University)!on!April!9,!2013.!6!For!further!information,!see!the!Land!Surface!Hydrology!Research!Group’s!webpage!at!http://hydrology.princeton.edu/data.pgf.php!(accessed!on!April!28,!2013).!

!

!

34!

normalized to a common resolution of 2 x 2 degrees by spatial averaging. The calculation procedures incorporate the non-uniform spacing of latitude bands employed in the gridding of each GCM. To include only the African continent, longitudes are bounded between 18 degrees west and 52 degrees east, and latitudes between 36 degrees south and 40 degrees north. In total, this creates a grid with dimensions 35 by 38 containing 1330 unique 2 x 2 degree grid cells.

STEP 3: APPLY SPATIAL D ISAGGREGATION

Next, LLNL approach is applied to spatially disaggregate the model outputs from 2 x 2 degrees to the resolution of the Princeton baseline dataset, or 0.5 x 0.5 degrees. This procedure applies to the 2 x 2 temperature and precipitation grid maps of Africa for each month (or day), year, and GCM. Inverse distance weighting is applied to increase the resolution of the modeled temperature and precipitation datasets to 0.5 x 0.5 degrees. Prior to bias correcting, all ocean grid cells are removed which reduces the 21,280 grid cells (i.e., 152 latitude grids x 140 longitude grids) to 10,639 land grids.

STEP 4: APPLY B IAS CORRECTION

With the GCMs and observed baseline gridded at a common 0.5 x 0.5 degree resolution, the next step is to bias correct the GCM outputs for each of the land grid cells. A baseline period of 1950 to 1999 is employed, and a projection period of 2001 to 2050. The three datasets used in the bias correction of each unique grid cell-GCM run combination for precipitation and temperature include:

1. Observed baseline between 1950 and 1999

2. Modeled baseline between 1950 and 1999

3. Modeled projections between 2001 and 2050

Using these three datasets, GCM bias correction is performed, taking the steps outlined below for each month (or day) and grid cell of each GCM run.

Step 4-1: Create a quantile map between the observed and modeled baselines. The quantile map includes the empirical cumulative density functions (CDF) of the observed and modeled baselines, which for the monthly procedure are the two 50-element datasets ranked and plotted side-by-side. For the bias correction of the daily datasets, the CDFs were made up of 50 years of 31 days each (1550 elements), where the 31 days included the day of interest and 15 days preceding and following. In so doing, it is assumed that the days in this 31-day rolling window are statistically similar, and ensure a much smoother CDF than if only 50 daily data points were included.

The quantile maps serve as the translators between each modeled projected value and its bias corrected equivalent. Given the chosen baseline, the quantile map is made up of two 50-year time series, and there is a unique quantile map for each month/day, GCM, grid cell, and meteorological variable (i.e., temperature and precipitation).

Step 4-2: Remove the temperature trend from the projected series. The bias correction process does not operate properly if the projected dataset is non-stationary. As a result, prior to applying bias correction, the temperature trend is removed from the 2001 to 2050 GCM outputs. Based on the approach employed by LLNL, the annual trend for each month was taken as the difference between a nine-year moving average of projected temperatures and the mean monthly baseline level.7 Following LLNL, there is assumed to be no precipitation trend through the 2050 period.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!7!Based!on!personal!communication!with!Edward!Maurer!(Associate!Professor!at!Santa!Clara!University)!on!April!9,!2013.!

!

!

35!

Step 4-3: Bias correct the projected precipitation and temperature series. For each year of the projected precipitation and temperature time series, the quantile map is used to create a bias corrected output. The general approach is to find the position of the projected value within the modeled baseline series, and then find the corresponding value in the observed series using the quantile map. That observed value is the bias corrected value. However, because not all projected values fall within the range of the modeled baseline series, the procedure differs depending on whether the projected value falls within or outside of the range of the modeled baseline time series.

• Within modeled baseline range. If the temperature or precipitation value falls within the modeled baseline range, then the non-exceedence probability of that value is located on the modeled baseline empirical CDF using interpolation, and the corresponding observed baseline value is taken at that non-exceedence probability as the bias corrected value.

• Outside modeled baseline range. If the value falls above or below the modeled baseline range, then extrapolating by fitting a distribution to the baseline modeled data, finds the location of the projected value on that theoretical distribution, and then maps that location to its equivalent on a distribution fit to the observed data. Following LLNL, for temperature values either above or below the modeled baseline, extrapolation is employed using a normal distribution. For precipitation values above the baseline range, a Gumbel distribution is fitted, and for values under the baseline range, a Weibull distribution is employed.

This procedure converts each of the projected temperature and precipitation values in the modeled projection time series to its bias corrected equivalent. Figure 1 demonstrates Steps 1 and 3 of the bias correction procedure. The graph on right shows the quantile map, comprised of CDFs of the observed and baseline modeled distributions for a given climate variable, grid cell, time period (month or day), and GCM. In the left graph, the solid black line is the raw GCM projection, and the dashed line shows the bias corrected output. The arrows connecting these two graphs show an example where the baseline GCM value is lower than the observed value for a given quantile, and this bias is then corrected for by upward-adjusting the corresponding projected GCM value.

FIGURE 1 B IAS CORRECTION PROCEDURE (SOURCE: LLNL)

Step 4-4: Add the temperature trend back into the bias corrected series. Finally, the temperature trend calculated in Step 4-2 is added back into the bias corrected series.

The outcome of this procedure is a set of monthly and daily CMIP5 climate projections for Africa that are bias corrected to be statistically consistent with observed historical data, and spatially disaggregated to a common resolution of 0.5 x 0.5 degrees.

!

!

36!

ANNEX(D(CHOICES(FOR(APPROACH(TO(CLIMATE(ANALYSIS(AND(UNCERTAINTY((The!preparation!of!the!Climate*Profile!required!making!a!number!of!methodological!and!presentational!choices.!!The!key!ones!are!explained!in!this!Annex.!!!!Why&do&we&use&Bias&Correction&and&Spatial&Disaggregation&to&downscale&GCM&outputs?&$Downscaling$GCMs!produce!output!at!various!scales,!with!spatial!grids!usually!ranging!from!a!1°!to!5°"spatial!resolution.!!At!these!variable!and!relatively!low!resolutions,!spatial!distinctions!and!variations!can!be!masked!(a!particular!risk!for!a!small!country!like!Uganda),!and!it!is!difficult!to!make!direct!comparisons!across!models.!!There!are!several!techniques!for!addressing!these!challenges.!!Three!commonly!used!techniques!are!dynamic!downscaling,!statistical!downscaling,!and!spatial!downscaling.!!The!latter!is!also!referred!to!as!spatial!disaggregation!and!is!the!technique!that!was!used!for!this!profile.!!!Each!of!these!techniques!is!briefly!summarized!below,!along!with!the!rationale!for!the!adopted!approach!of!spatial!disaggregation.!!Dynamic*Downscaling*can!be!achieved!in!two!ways:!• General!Circulation!Model!(GCM)!output!is!applied!as!boundary!conditions!to!a!regional!climate!model!

(RCM).!!RCMs!have!a!finer!resolution!than!coarser!GCMs,!with!more!detail!on,!e.g.,!!land!surface!or!cloud!coverage.!!Thus,!they!are!able!to!produce!results!at!a!higher!resolution.!

• A!variable!resolution!global!model!(VRGM),!where!a!higher!resolution!grid!is!embedded!into!a!GCM,!can!also!be!used.!!This!allows!the!GCM!to!produce!results!of!various!resolutions,!as!well!as!detailed!spatial!results!for!certain!regions!of!interest.!!!

!Dynamic!downscaling!requires!either!RCMs!or!VRGMs,!access!to!the!underlying!GCM,!and!large!amounts!of!computational!capacity,!making!it!infeasible!to!reproduce!for!the!numerous!GCMs!employed!in!the!profile.!!!!Statistical*Downscaling*is!a!method!that!develops!statistical!relationships!between!climatic!variables!and!regional!variables.!!It!is!computationally!less!demanding!than!dynamic!downscaling!and!can!often!effectively!characterize!modeling!uncertainties.!!However,!with!an!incomplete!or!inaccurate!regional!climate!dataset,!statistical!downscaling!can!be!difficult!to!achieve,!or!can!be!misleading!in!its!perceived!accuracy.!!!!Spatial*downscaling,!more!accurately!described!as!spatial!disaggregation,*is!the!most!commonly!used!form!of!downscaling.!!It!is!a!mathematical!‘smoothing’!technique!that!uses!inverse!distance!weighting!or!another!method!to!disaggregate!the!GCM!data!to!a!higher!resolution.!!In!the!context!of!this!profile,!once!disaggregated,!the!higher!resolution!GCM!data!was!bias!corrected!using!a!more!spatially!refined!baseline!(the!Princeton!dataset,!discussed!below).!!!!!!!Spatial!downscaling!was!used!for!this!profile!because!a!primary!aim!of!the!analysis!was!to!obtain!results!from!multiple!models!in!order!to!capture!the!full!range!of!possible!climate!futures!(also!refer!below).!!Applying!more!complex!downscaling!techniques!to!the!ensemble!of!projections!would!have!been!prohibitively!costly!in!terms!of!time!and!other!resources.!!In!addition,!this!approach!produces!high!resolution!results!while!essentially!maintaining!the!GCMs!in!their!“raw”!state,!thus!avoiding!concerns!about!potential!bias!introduced!by!more!sophisticated!downscaling!procedures.!!Further!details!on!the!

!

!

37!

methodology!used!in!this!analysis!to!produce!Spatially:Disaggregated!and!Bias!Corrected!(SDBC)!climate!projections!are!provided!in!Annex!C.!!Why&do&we&report&the&results&from&several&models?&!Climate!modeling!is!inherently!complex!and!some!models!are!believed!to!be!more!robust!than!others.!!However,!there!is!no!agreed!upon!method!for!rejecting!any!one!model’s!validity.!!Because!outputs!from!various!models!can!differ!quite!significantly,!presenting!projections!from!only!one!model!can!be!misleading.!For!these!reasons!and!in!order!to!capture!the!full!range!of!possible!future!climate!outcomes,!the!IPCC!recommends!using!a!large!set!of!GCM!projections.!!As!shown!in!this!profile,!there!are!some!outcomes!on!which!models!tend!to!agree;!for!other!outcomes!there!is!less!model!agreement!and!more!uncertainty.!!It!is!important!to!understand!these!patterns!of!agreement!and!uncertainty,!but!they!are!less!apparent!when!using!a!smaller!set!of!GCMs.!!!!Why&do&we&consider&only&two&emissions&scenarios?&&

In!the!latest!Fifth!Assessment!Report!(AR5)!of!the!IPCC,!four!representative!concentration!pathways!(RCPs)!are!presented:!RCP!2.6,!RCP!4.5,!RCP!6!and!RCP!8.5.!!RCPs!are!defined!by!their!level!of!radiative!forcing!relative!to!1750!levels!(in!W/m2).!!!The!lowest!warming!scenario,!RCP!2.6,!is!considered!to!be!highly!unlikely!at!this!point.!!Therefore,!in!order!to!capture!the!likely!range!of!RCP!results,!the!more!plausible!lower!end!scenario!(RCP!4.5)!and!the!high:end!scenario!(RCP!8.5)!are!used.!!(!In!previous!IPCC!reports,!scenarios!from!the!Special!Report!on!Emissions!Scenarios!(SRES)!are!used!to!describe!potential!future!scenarios.!!Instead!of!being!defined!by!radiative!forcing,!SRES!scenarios!are!defined!by!various!economic!and!social!development!pathways!that!act!as!the!driving!force!for!climate!outcomes.!!Figure"D1"from!Rogelj!et!al!(2012)!compares!the!globally!averaged!temperature!projections!from!SRES!scenarios!and!RCPs.!!!

!Previous!work!on!climate!change!conducted!for!the!Uganda!Country!Water!Assistance!Strategy!used!two!SRES!scenarios,!namely!the!A1B!and!A2,!with!GCMs!from!the!CMIP3!ensemble.!!!Some!broad!comparisons!between!these!two!SRES!scenarios!(using!CMIP3!GCM!runs)!and!the!two!RCPs!used!in!this!profile!(using!CMIP5!GCM!runs)!can!be!drawn!from!Figure"D2"and"D3,"which!illustrate!ensemble!projections!averaged!over!Uganda.""!First,!note!in!Figure"D2"that!RCP!4.5!and!RCP!8.5!envelope!the!temperature!ranges!projected!from!the!SRES!scenarios!from!CMIP3.!!!Also!note!in!Figure"D3!that!the!range!in!projections!for!

annual!precipitation!is!greater!for!the!SRES!scenarios!than!for!the!RCP!scenarios,!though!the!median!model!projections!remain!comparable.!!!In!other!words,!the!use!of!the!two!RCPs!broadly!captures!the!ranges!of!the!SRES!scenarios.!!!Why&do&we&report&results&at&a&WMZ&level&and&map&projections&on&a&0.5o&x&0.5o&scale?&&

Figure&D1:&Figure(3b(from(Rogelj(et(al((2012),(temperature(change(projections(of(SRES(and(RCP(scenarios(

!

!

38!

Climate!and!climate!change!projections!can!be!reported!on!various!spatial!scales.!!In!this!profile,!GCMs!were!spatially!downscaled!to!a!0.5°!x!0.5°!resolution!to!match!the!resolution!of!the!Princeton!historical!dataset.!!Mapping!results!at!this!scale!and!reporting!on!model!agreement!provides!insight!into!projected!patterns!of!climate!change!across!Uganda.!!However,!conclusions!drawn!about!climate!projections!at!this!scale!could!be!erroneous.!!Since!GCMs!are!run!at!a!coarser!resolution,!typically!ranging!between!1°!and!5°!,!reporting!results!at!a!0.5°!x!0.5°!resolution,!while!more!precise,!is!a!less!accurate!representation!of!GCM!projections.!!!At!the!same!time,!spatially!distinctive!climate!projections!can!be!lost!when!reporting!results!only!on!a!higher!scale,!e.g.,!at!a!country:wide!level.!!For!instance,!if!the!north!is!expected!to!become!dryer!and!the!south!wetter,!averaging!across!the!country!could!show!no!projected!changes!in!aridity,!which!is!an!incorrect!conclusion!to!draw.!!The!approach!taken!in!this!profile!–!to!report!results!at!the!WMZ!level!(in!addition!to!country:wide),!while!mapping!projections!on!a!0.5!x!0.5o(scale(–!strikes!a!balance!between!‘overly’!coarse!and!‘overly’!fine,!lower!and!higher!resolutions.!!!!!Reporting!at!the!WMZ!level!is!appropriate!for!two!other!important!reasons:!First,!!climate!and!hydrologic!systems!are!closely!inter:connected;!climate!is!a!driving!force!behind!dynamic!hydrologic!processes!and!conditions,!and!vice:versa.!!!Second,!the!natural!hydrologic!unit!is!a!catchment!and!Uganda!is!currently!following!catchment!boundaries!–!the!WMZs!–!for!water!management!and!development,!making!information!at!this!level!of!practical!use.!!!!Why&do&we&use&the&Princeton&dataset?&&&Historical!data!for!Uganda!was!required!for!the!analysis,!both!to!characterize!historical!climate!and!to!calibrate!model!projections!to!historical!data!for!the!region.!!Because!regional!climate!is!defined!by!the!means!and!variability!of!certain!properties!over!an!extended!period!of!time,!the!characterization!of!historical!climate!and!the!calibration!of!climate!models!is!most!accurate!using!historical!data!that!span!multiple!decades.!!!Ideally,!reliable!station!data!would!be!available!at!multiple!stations!scattered!throughout!the!country!over!extended!and!uninterrupted!period!of!time.!!Station!data!for!Uganda!are!not!available!with!this!spatial!or!temporal!coverage,!so!the!use!of!station!data!was!not!an!option.!!!

Figure&D2:(Temperature(projections(averaged(across(Uganda(for(RCP(and(SRES(scenarios((

Figure&D3:(Temperature(projections(averaged(across(Uganda(for(RCP(and(SRES(scenarios((

!

!

39!

An!alternative!approach!for!obtaining!historical!climate!is!to!use!global!datasets!that!provide!estimates!of!gridded!historical!climate!combining!global!observation:based!data!with!reanalysis!data.8!!The!Climate!Research!Unit!(CRU)!dataset!from!the!University!of!East!Anglia!is!arguably!the!most!cited!global!historical!dataset.!!The!data!are!gridded!at!a!0.5°!x!0.5°!resolution,!but!are!available!only!on!a!monthly!time:step.!!Capturing!acute!climate!extremes!requires!daily!data,!ruling!out!the!possibility!of!using!the!CRU!dataset!for!this!profile.!!Instead,!the!Princeton!land!surface!dataset!was!used.!!It!is!an!equally!reputable!historical!dataset!available!for!daily!time!steps!at!resolution!of!0.5°!x!0.5°.!!A!comparison!of!Princeton!data!with!station!data!was!undertaken!in!order!to!establish!the!extent!to!which!the!two!approximate!one!another!(see!Annex!E).!&Why&do&we&report&only&on&CMIP5?&(This!profile!reports!only!on!CMIP5!runs!for!a!few!key!reasons.!!!First,!the!two!RCPs!used!envelope!the!SRES!scenario!temperature!projections,!providing!a!sufficiently!broad!range!of!outcomes!(refer!above).!!Second,!the!profile!reports!on!a!large!volume!of!data!–!including!from!approximately!30!GCMs,!two!time!frames,!and!two!RCP!scenarios!–!and!reports!on!nearly!20!climate!indicators.!!Including!yet!more!scenarios!and!models!would!make!the!results!exceedingly!difficult!to!understand!without!necessarily!providing!more!information.!!Finally,!selecting!from!the!CMIP5!ensemble!is!appropriate!as!the!ensemble!provides!the!most!up!to!date!GCM!runs!using!the!latest!agreed!climate!pathways!(i.e.!RCPs).!!!!

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!8!Reanalysis!is!a!method!that!combines!observed!data!with!a!numerical!weather!prediction!model!to!generate!a!gridded!historical!climate!dataset.!!!!

!

!

40!

ANNEX(E(Comparison"between"Available"Station"Data"and"Princeton"Data"!This!Annex!provides!a!summary!of!the!statistical!comparison!between!available!station!data!and!the!Princeton!dataset!used!for!the!analysis!presented!in!this!report.!!As!discussed!in!Annex!D,!performing!the!analysis!for!the!report!required!data!at!a!high!resolution!(0.5!x!0.5!degrees)!over!a!period!of!several!decades.!!!Due!to!an!insufficient!number!of!stations!and!the!short!time!period!spanned!by!them,!monitoring!stations!were!unable!to!provide!sufficient!data!for!the!analysis.!!Therefore,!the!Princeton!dataset!was!used!instead.!!In!order!to!ensure!compatibility!between!the!Princeton!dataset!and!available!station!data,!a!brief!statistical!comparison!of!these!two!datasets!was!conducted.!!!Of!the!14!stations!with!available!station!data,!10!of!them!have!data!that!cover!a!24!year!span!(1990:2013).!!The!other!4!stations!are!missing!several!years!of!data!and!so!were!not!included!in!the!comparison.!!!The!Princeton!dataset!spans!the!period!of!1948!to!2008.!!There!are,!therefore,!19!years!of!overlapping!data!(1990:2008)!between!the!two!datasets.!!!Station!data!include!monthly!precipitation,!mean!monthly!maximum!temperature,!and!mean!monthly!minimum!temperature.!!The!comparison!was!conducted!for!these!three!variables.!!!Station!locations!were!approximated!and!the!‘matching’!Princeton!data!were!obtained!using!inverse!distance!weighting!of!the!four!grid!points!surrounding!the!station!data!location.!!The!comparisons!were!made!using!monthly!values!of!the!three!variables!over!the!19:year!period!for!the!approximated!station!locations.!!!!For!precipitation,!Princeton!data!and!station!data!patterns!match!quite!closely!(See!Figure"E1"and"Table"E1).!!However,!Princeton!data!generally!do!not!capture!the!same!extremes!that!are!reached!at!the!stations.!!This!is!not!surprising!considering!that!Princeton!data!are!averaged!over!an!entire!grid!cell,!thereby!muting!local!extremes.!!While!temperature!patterns!are!matched!well!(Figures"E2"and"E3,"and"Tables"E2"and!E3),!there!is!some!discrepancy!in!temperature!averages:!Princeton!data!are!either!consistently!higher!or!consistently!lower!than!station!data.!!This,!again,!is!not!surprising!since!Princeton!data!represent!the!average!temperature!of!a!grid!cell,!and!so!mute!temperature!variation!that!occurs!due!to!altitude!changes.!!For!example,!if!station!data!are!recorded!at!a!higher!than!grid:cell!average!altitude,!temperatures!can!be!expected!to!be!consistently!lower!than!the!grid:cell!average.!!!There!are!some!stations!where!the!discrepancy!between!station!data!and!Princeton!data!is!most!likely!due!to!errors!in!station!data.!This!is!the!case,!for!example,!for!Lira!station!mean!monthly!temperature.!!

! !

!

!

41!

! !!!!

! !

! !

! !!Figure"E1:"Monthly"precipitation"""

!

!

42!

"Figure"E2:"Mean"Maximum"Monthly"Temperature""!!

!Figure"E3:"Mean"Minimum"Monthly"Temperature"!!For!all!three!variables!and!each!station,!T:tests!and!Kolmogorov:Smirnov!tests!were!used!to!test!whether!Princeton!data!and!station!data!follow!the!same!or!different!distributions.!!Both!tests!indicate!that!there!is!sufficient!evidence!(at!95%!confidence)!that!only!3!of!the!10!stations!do!not!come!from!the!same!distribution!as!the!Princeton!data!for!precipitation.!!Both!tests!rejected!the!Null!Hypothesis!(that!station!data!and!Princeton!data!come!from!the!same!distribution)!for!all!of!the!stations!for!both!maximum!and!minimum!monthly!temperatures.!!This!is!a!result!of!mean!temperature!differences!between!station!data!and!Princeton!data!(see!Table"E1,"Table"E2"and!Table"E3).!!!!Although!there!is!insufficient!station!data!in!terms!of!spatial!resolution!and!length!of!time!to!allow!a!complete!comparison!between!station!data!and!Princeton!data,!the!results!show!that!Princeton!data!

!

!

43!

are!a!close!enough!approximation!to!provide!confidence!in!the!analysis!presented!in!the!Climate!Profile.!!!!Table"E1"Station"Name"

Precipitation""(mm/month)""Mean" Standard"Deviation" t^test""(α"=0.05")"

Null"Hypotheses:""Station" Princeton" Station" Princeton"Gulu" 121.7" 112.5" 84.9" 77.8" Do"Not"Reject"Lira" 123.7" 104.9" 81.8" 68.3" Reject"Masindi" 109.8" 107.0" 70.2" 65.5" Do"Not"Reject"Soroti" 113.2" 100.2" 76.4" 66.1" Do"Not"Reject"Tororo" 128.7" 116.5" 71.9" 78.2" Do"Not"Reject"Jinja" 111.5" 119.9" 69.6" 72.0" Do"Not"Reject"Kasese" 71.1" 92.0" 48.2" 52.4" Reject"Mbarara" 78.9" 86.7" 53.2" 52.0" Do"Not"Reject"Entebbe" 132.2" 117.9" 102.6" 70.7" Reject"Kabale" 85.0" 90.0" 52.0" 53.3" Do"Not"Reject"

!!!!Table"E2"Station"Name"

Mean"Monthly"Maximum"Temp"(°C)"Mean" Standard"Deviation" t^test""Station" Princeton" Station" Princeton"

Gulu" 30.2" 31.7" 1.89" 2.10" Reject"Lira" 30.6" 31.7" 1.84" 2.15" Reject"Masindi" 28.8" 30.6" 1.78" 1.72" Reject"Soroti" 30.4" 32.2" 1.77" 2.02" Reject"Tororo" 29.7" 30.7" 1.41" 1.63" Reject"Jinja" 28.2" 30.5" 1.14" 1.47" Reject"Kasese" 30.6" 28.1" 1.14" 1.24" Reject"Mbarara" 27.3" 29.9" 1.34" 1.23" Reject"Entebbe" 26.5" 29.6" 1.29" 1.31" Reject"Kabale" 24.1" 25.5" 0.81" 1.15" Reject"

!Table"E3"Station"Name"

Mean"Monthly"Minimum"Temperature"Mean" Standard"Deviation" t^test""Station" Princeton" Station" Princeton"

Gulu" 18.5" 18.8" 0.82" 1.25" Reject"Lira" 15.2" 18.3" 2.35" 1.13" Reject"Masindi" 17.7" 18.7" 1.02" 1.16" Reject"Soroti" 18.4" 19.1" 0.77" 1.10" Reject"Tororo" 17.1" 16.6" 0.86" 1.11" Reject"Jinja" 17.1" 17.8" 0.72" 1.12" Reject"Kasese" 17.8" 16.0" 0.78" 1.16" Reject"Mbarara" 15.3" 14.7" 0.89" 1.18" Reject"Entebbe" 18.2" 17.3" 0.83" 1.16" Reject"

!

!

44!

Kabale" 11.4" 12.6" 1.16" 1.18" Reject"!

Documents

Climate Profile May27 - Makerere Universitymuccri.mak.ac.ug/sites/default/files/Publications/Climate... · 2016-04-07 · iii! (ACKNOWLEDGMENTS( This!report!was!prepared!by!Megan!Lickley!(Climate!Specialist,!World!Bank!consultant),!supported!by!Halla!