The desiccation of southern Africa’s Okavango Delta: Periodic fl uctuation or long-term trend?

HAMISAI HAMANDAWANA Department of Environmental Science, University of Botswana; [email protected]

The Okavango Delta is the largest inland delta on Earth, spreading over 15,000 km2 of the Kalahari sands of Botswana. Though commonly referred to as a delta, it is actu- ally a landlocked wet-fan comprising three active subsystems: permanent, seasonal �� Special Section: Highlights Science and intermittent fl oodplains bounded by fossil fl oodplains that mark the cotermi- nous extent of a more extensive wetland during the historical past (Fig. 1). Contem- porary perspectives on climate change in this sub-region are dominated by two competing discourses. The fi rst maintains that the current trend in increasing aridity is part of periodic variation within a stable equilibrium and need not be interpreted as sustained deterioration of climatic con- ditions. The second argues that though pe- riodic variation is evident, close examina- tion of this periodicity reveals a persistent downward trend; a trend that may imply only limited prospects for recovery to the wetter climatic conditions of the historical past. Recent long-term evidence compiled from disparate sources about this sub-re- Figure 1: a) PPositionosition ooff llocationsocations mmentionedentioned iinn ttextext aandnd eextentxtent ooff tthehe OOkavangokavango DDelta’selta’s mmajorajor ssubsystems;ubsystems; b) a gion’s climate since the beginning of the sketch drawing of David Livingstone (early missionary explorer) and his family on the shores of Lake Ngami in 1849; c) LakeLake Ngami’sNgami’s surfacesurface waterwater distributiondistribution inin 1967,1967, 19891989 andand 20012001 asas reconstructedreconstructed fromfrom historicalhistorical recordsrecords 19th century suggests that there is need and CORONA and Landsat imagery. For the graph in insert 1c, fi gures for 1849 are based on Livingstone’s lower to question the validity of the stable-equi- estimate of the Lake’s perimeter (70 km) in 1849. librium hypothesis. In this sub-region, oral histories from vations are indicative of a gradual decline reed swamps infested by buff aloes and local people off er an invaluable form of in rainfall since the early 19th century: elephants that had to be hunted from evidence about climate conditions during the drying up of the Selinda spillway dur- boats (Chapman, 1886)) last fl owed into the historical past. Around 1800 AD, the ing the early 1870’s (Stigarnd, 1922), Lake Lake Ngami around 1883 (Stigarnd, 1922). baYeyi migrated from southern Zambia by Ngami’s gradual desiccation (Fig. 1b), and Since then it has retreated northward to its canoeing along the Selinda spillway and virtual disappearance of peat fi res in this outfl ow point from the Delta’s permanent distributary channels of the Okavango environment by the late 1920’s, all confi rm swamps by the late 1960’s (Fig. 1a). Given River to settle around Lake Ngami (Fig. a continuous drying trend (Hamandawana the protracted failure of these emergent 1a) where they had been attracted by the et al, 2005). fl oodplain environments to periodically abundant presence of hippo and fi sh. This Though the Lake Ngami’s fl ood ré- recover, as suggested by proponents of information suggests high fl oods in the gime has been characterized by periodic quasi-periodic oscillations, it is quite ap- Okavango Delta that were capable of sus- fl uctuations, the long-term trend suggests parent that this sub-region’s climate has taining perennial outfl ow into the Selinda that the hydrological system has been un- become increasingly drier from the begin- spillway and permanent water presence able to revert to the higher fl ood condi- ning of the 19th century to the present. in Lake Ngami during the fi rst half of the tions of the historic past. This proposition Though human interventions have con- 19th century. The latter is corroborated is supported by simultaneous fl oodplain tributed to fl oodplain desiccation (i.e. pa- by archival evidence, with the presence of desiccation in peripheries of the present pyrus rafts causing vegetation blockages fi shing boats in the background of Figure Delta’s permanent and seasonal swamps. and channel manipulations by colonial 1b pointing to perennial water residence For example, in the east, the Mababe De- authorities), evidence suggests that the and a productive environment. High pro- pression (captured in oral traditions as sustained decrease in rainfall is the major ductivity is further confi rmed by docu- Lake Mababe) dried during the second cause of persistent fl oodplain contraction mented occurrence of widespread peat half of the 19th century; its infl ow chan- (Hamandawana et al., in press). fi res in this area during the early 1920’s, nels from the northwest and southwest While the foregoing examples of oral which suggest substantial papyrus growth ceasing to fl ow by the early 1960’s (Camp- and documentary evidence provide valu- capable of sustaining the accumulation of bell and Child, 1971). In the west, the Tha- able indicators of the general direction of considerable peat deposits. These obser- oge River (described during the 1880’s as climate change, additional corroborative 12

PAGES News, Vol.15 • No 1 • April 2007 Special Section: Highlights Science Figure 2: Frequency of arid years by decade period for Shakawe and Maun: 1955-2004. evidence comes from trends in mapped in local rainfall, which has tended to se- designed to enhance human capacities to vegetation distribution during the second lectively facilitate the expansion of woody cope with deteriorating climate conditions half of the 20th century. In the proximal, in- cover at the expense of drought-sensitive are urgently required. Adoption of eff ec- termediate and distal reaches of the Oka- wetland and dryland grasses. Though this tive strategies requires offi cial acknowl- vango Delta, recent changes in vegetation area’s climate is widely classifi ed as semi- edgement of the non-transient character distribution have been characterized by arid ((Aridity Index (AI) = 0.20-<0.50)), of the present direction of change. Thus transitions from reed-swamps and open analysis of available data shows increasing the onus is on climate change science to grassland to woody cover dominated by frequency of arid years (AI = 0.05-<0.20) inform policy formulation by providing drought tolerant species (Hamandawana, for successive decades between 1955 and long-term perspectives. 2006). Between 1967 and 2001, woody 2004 (Fig. 2). cover around Shakawe and Mohembo This shift from semi-arid to arid con- References increased by 20% while open grassland ditions is consistent with current climate Campbell, A. and Child, G., 1971: The impact of man on the environment of Botswana, Botswana Notes and Records, 3: 91-109. declined by 15% with these trends being change scenarios that point to mid-conti- Chapman. J., 1886: Travels in the interior of South Africa, Bell and Daldy: statistically signifi cant at α = 0.01 and 0.05 nent drying in southern Africa, centered on Reprinted A.A. Balkema, Cape Town, 1971. respectively. In the intermediate reaches Botswana, due to the increasing incidence Hamandawana, H., Eckardt, F. and Chanda, R., 2005: Linking archival and around Gumare, woody cover signifi cantly of rainfall failures. In view of the pervasive remotely sensed data for long-term environmental monitoring, International Journal of Applied Earth Observation and Geoinfor- increased by 26% (α = 0.05) as open grass- nature of drying sequences during the re- mation, 7(4): 284-298. land declined by 43% (α = 0.01). In the cent past, the need for formulating appro- Hamandawana, H., 2006: An investigation of the role of human factors in the degradation of natural resources in and around the Okavango distal reaches around Sehitwa, Acacia mel- priate policies designed to mitigate the Delta, Ph.D. Thesis, Department of Environmental Science, Uni- lifera increased by 21% while open grass- adverse eff ects of deteriorating climatic versity of Botswana. land declined by 5% with both trends conditions is now overdue. With evidence Hamandawana, H., Chanda, R. and Eckardt, F., in press. Reappraisal of contemporary perspectives on climate change in southern Afri- being statistically signifi cant at α = 0.01. suggesting that these sub-regional trends ca’s Okavango Delta sub-region, Journal of Arid Environments. These changes are indicative of climate- are likely to persist, deployment of appro- Stigarnd, A. G., 1922: Sketch maps of Ngamiland and Ghanzi: Plan BP-122 driven responses to progressive decrease priately informed adaptation strategies and BP-123. Department of Surveys and Mapping, Gaborone.

Sensitivity of wetlands and water resources in south- eastern Australia to climate and catchment change

PETER GELL1, R. JONES2 AND A. MACGREGOR1 1Geographical & Environmental Studies, University of Adelaide, Australia; [email protected] 2CSIRO Marine & Atmospheric Research, Aspendale, Australia. Southeastern Australia lies within a tem- rapid decrease in moisture at 5.5 kyr BP fol- ple, along the River Murray, Tareena Billa- perate zone characterized by mild, but lowed by a gradual trend to a variable and bong experiences mid-Holocene freshness wet winters and hot summers with low drier climate at c. 3 kyr BP. From c. 2.2 kyr BP followed by increased connectivity to the ratios of precipitation:evaporation (P/E). the lakes refi lled, though not to the mid- River (Gell et al., 2005) but then a sustained Reconstructions of paleosalinity and water Holocene levels. These lakes are therefore period of stability until the arrival of Euro- level from numerous crater lakes (e.g. Lake sensitive to changes in P/E on the scale of pean settlers c. 1840 AD. In the Coorong, a Keilambete) provide the means to moni- centuries to millennia but appear insensi- large back-barrier lagoon at the mouth of tor long-term changes to P/E, and have tive to short-term climate variability. In the River Murray system, stratigraphic and revealed a regional pattern of Holocene contrast, the diatom-inferred paleosalin- diatom evidence reveals relative resilience climate change (Bowler, 1981; Chivas et al., ity records from numerous fl uviatile lakes (Fluin et al., in press) to the P/E changes 1985; Gell, 1998; Jones et al., 1998). These across southeastern Australia across this documented in the sensitive crater lakes records witness humid conditions and period do not reveal the same sensitivity to the east. Still further east in the coastal overfl owing lakes in the mid-Holocene, a to long-term changes in climate. For exam- plain of the Snowy River, a diatom salin- 13

PAGES News, Vol.15 • No 1 • April 2007