42 Erdkunde Band 41/1987

THE DEVELOPMENT OF THE STRZELECKI DUNEFIELDS, CENTRAL AUSTRALIA

With 4 figuresand 2 tables

Ernst Loffler and Marjorie E. Sullivan

Zusammenfassung: Zur Entwicklung der Diinenfelder in Although the inherited pattern recognized suggested der Zentralaustralien - - Strzelecki-Wiiste, a former enlarged lake Lake Dieri no supporting In einem friiheren Aufsatz die dafi postulierten Autoren, field evidence was available. von grofie Gebiete der Strzelecki- und Simpson-Wiisten a - - The existence of formerLake Dieri has since been einem Riesensee Lake Dieri bedeckt gewesen waren und accepted by subsequent workers (Wasson 1983 a, b) dafi die regelmafiigen Langsdiinen, die heute die Wusten but there is disagreement as to the development of the iiberziehen, entstanden, als sich Lake Dieri aufgrund fort which we saw as linked schreitender Austrocknung langsam zuriickzog. Der Sand, longitidinal directly von with the of Lake Dieri. We that the der die heutigen Diinen aufbaut, soil grofitenteils ehe history argued dunes which extend from the maligen Stranddiinen stammen, die im Lee des Riesensees regular longitudinal entstanden waren und die ihrerseits ihr Material von chain of salt lakes in a fan-like pattern developed as den von Osten und Nordosten her kommenden Fliissen er Lake Dieri receded, and that the sand originated hielten. largely from littoral dunes which formed along the lee Zur der Lake Dieri wurde eine Verifizierung Hypothese side of the lake.With the gradual retraction of the lake Reihe von Sandproben entlang zweier Traversen ostlich shores, new sets of littoral dunes were formed and der Salzseen und Lake Callabonna gesammelt sand from the older shoreline dunes was blown out und Die dafi des Diinen analysiert. Analysen zeigen, langs and into the felds weder deutliche Unterschiede in der mittleren Korn incorporated logitudinal system. tests is es One of the crucial of this interpretation the grb&ensortierung bestehen, noch dafi deudiche Hinweise relative of the sand across the dunefield. If it can in der Komgro&enverteilung gibt, die auf unterschiedliche age be established that the dune from Lange des Sandtransports schliefien lassen. Hingegen be system developed steht ein deutlicher Abfall im Feldspatgehalt mit steigender outside in, i.e. that sand within the dunes furthest von Entfernung den heutigen Salzseen. Diese Tatsachen away from the present lake system is oldest and that von zeigen, dafi erstens der Sand einem relativ einheitlichen sand within those nearest to the lake chain is youngest dafi zweitens der Sand nicht iiber das Liefergebiet stammt, the earlier interpretation would be supported. For wurde und dafi gesamte Diinenfeld aolisch transportiert reason a was on this number of sand samples collected = an drittens der Grad der Verwitterung ( Verlust Feldspat) two traverses across the Strzelecki Desert (Fig. 2) and und damit das relative Alter der Diinen mit grofier werden mechanical and chemical were carried out von analyses der Entfernung den heutigen Salzseen zunimmt. Das on the sands to characterize them and to any bedeutet aber, dafi sich die Diinen von Osten und Nord identify trendswhich have on their or osten her entwickelten in Ubereinstimmung mit dem Lake may bearing age origin. The two traverses followed the two tracks Dieri Konzept. only from west to east and southwest crossing the desert to northeast respectively. Traverse 1 followed the Strzelecki Track fromLake Callabonna to Innamincka Introduction and traverse 2 a track from Lake Frome toHawker Gate (Fig. 2). Because of the proximity of the Strze we In an earlier paper (Loffler a. Sullivan 1978) leckiCreek with its considerable sand load in the first was suggested from an interpretation of satellite imagery traverse itwas difficult to collect sand that not that a former lake occupied much of the area now influenced by recent sand transport from the creek covered by the dunefields of the Strzelecki and bed and the samples could therefore not be collected Simpson , its legacy being series of pans in fixed intervals as on the second traverse where two were ten kilometres. Where aligned more or less concentrically and following the sand samples taken every were from the dune trend of the present eastern and northern shorelines of possible surface samples taken the chain of salt lakes between Lake Eyre in the crestwhere the sand is loose (Samples A) and from the northwest and Lake Frome in the southeast (Fig. 1). indurated sand at the toe of the dune (Samples B). Ernst Loffler andMarjorie E. Sullivan: The development of theStrzelecki Desert dunefields, Central Australia 43

^ J*

i: Landsat Fig. image (quarter scene) of northernmost part of Lake Frome and Lake Callabonna (top) and the pattern of aligned pans and northeasterly trending longitudinal dunes east of the lakes. The alignment of the pans transverse to dunes accentuated are water on longitudinal is because the pans filled with after heavy rain and show up clearly the infrared band of the Landsat imagery

Landsatszene (lA Ausschnitt) mit nordlichem Abschnitt des Lake Frome und Lake Callabonna und dem Muster der nebeneinander gereihten Tonpfannen und Langsdunen ostlich der Seen. Das Muster der Tonpfannen wird dadurch zur hervorgehoben, dafi die Tonpfannen Zeit der Aufnahme wassergefiillt waren, was sich besonders deutlich auf dem infraroten Kanal (Band 7) zeigt 44 Erdkunde Band 41/1987

Birdsville^"

^^^^^^l^ I nnarr^

~~ """"^ Trendof longitudinaldunes I W.A r??^\ ' ^^o/^ '"i ^ ! of S.A.^H^-Trend^ alignedpans /^^i , N.SW/ Former drainage lines ---^^ \^?^ ... ^Z^a/ce ^ggy_J LocalityMap \|vic~:\/ ^J\S^ ^< High ground<^>^ liKr 0_100km

2: Locality map with position of traverses

Ubersichtskarte mit Angabe der ungefahren Lage der Traversen

as an in a. Sand analyses been interpreted increase age (Wopfner Twidale 1967). Initially three methods were used Mechanical analyses of the sand were carried out on a number of selected samples of traverse 1 to by dry sieving duplicate 100 g samples through a set determine the feldspar/quartz ratios in order to most of nested Endecott test sieves at 0,5 phi intervals. establish the economic working method. Cumulative percentage weight values were plotted, The threemethods are: sorting and skewness parameters calculated according 1. Analyzing the sand forAl, Na and K and calcu amount A to themethod described by Folk (1974), and descrip lating the of feldspar. representative tions of the sand samples made using Wentworth size sample was agitated inwater with sodium hydro classes (Folk 1974). These analyses were carried out xide and calgon to disperse the sediment and to characterize the sands and to detect any major remove clay from the sand grains. The clean sand changes which may reflect the origin of the sediment was then submitted for chemical analysis to deter or themethod of its deposition. mine the proportions of the feldspar cations. Optical and chemical analyses were carried out on 2. Making a thin section from a resin-impregnated small subsamples of these sands to determine their block of clean sand and counting the proportion on contained proportion of feldspar and quartz, and the of quartz and feldspars in 500 grains the no total amounts of iron (both ferric and ferrous). This assumption that there is systematic difference or was done to detect any chemical changes variations in grain size between quartz and feldspar sand some on in the sands that may have bearing the grains. a relative ages of sands across the dunefield. In particu 3. Preparing resin impregnated block of clean lar the increase in reddening across the dunefield has sand, staining the surface using hydrofluoric acid Ernst E. Lqffler andMarjorie Sullivan: The development of theStrzelecki Desert dunefields, Central Australia 45

and sodium and Cum Cum . cobaltnitrate, point counting for 100-1% a D 100-j% ^ quartz, potassium feldspar and plagioclase. The results of the analyses are shown inTable 1. The results of all threemethods were similar in so far as theywere internally consistent however the total amount of feldspar varied between methods with the 50- /50- / total amount the feldspars being considerably lower inmethod 2 than inmethods 1 and 3. The very low count of feldspar grains inmethod 2 is partly due to the difficulties of differentiating between and Grains 0-1-1-1-1-1 ? 0-1-^?T~-T-1 microscopically quartz feldspar. 0 1 2 3 4? 0 1 2 3 4? below 30 to 40 fim in size were not counted because of Fig. 3: Typical examples of curve of upper and the increased difficulties of identification, due in sorting (a) lower dune sand to (b) particular the decreased chance of characteristic Typische Beispiele der Sortierungskurve eines Diinen feldspar twinning being displayed in a small area. sandes. a: Sand vom b: Sand von der Basis Also have been mis-identified Dunenkamm, feldspar grains may der Dime as in cases are quartz where they untwinned or oriented not to show twinning, or do not show signs of or cleavage alteration. An underestimate of the amount of feldspar is therefore likely. The chemical analyses show that there is no Methods 1 and 3 however produced similar results apparent trend in the iron content (both ferric and even the exact amount though of feldsparmay differ. ferrous) with increasing distance from the lakes these twomethods were used for Only analyzing the (Tables 1, 2), despite an obvious change in colour collected at a later date traverse samples along 2.We from pale yellow to bright red across the dune field. consider method 1 to be themost satisfactory because This is in agreement with general observations that itproduced themost consistent results and is simplest the amount of iron is no measure for redness in dune and cheapest. sands (Gardner a. Pye 1981). Wasson's (1983 b) results show a slight increase in iron content from west to east, but asWasson points out his sample is Results themechanical and chemical too of analyses small to draw any conclusions. Only two samples of his traverse show higher iron content. Our results Mechanical analyses of the dry sand indicated also do not support his conclusion that the red sands a uniform of sand a relatively body throughout the have higher iron content than the pale sands. If area. The sand was to sampled moderately modera anything there is a slightlyhigher iron content in the well fine skewed fine tomedium sand. near tely sorted, pale sands the salt lakes (Tables 1,2). Local of coarse examples skewed sand occurred, The lack of any systematic differences in modal from near a more generally pan surfaces where grain size sorting across the dunefield and the relative was gravelly lag concentrated. consistant Fe content indicate that the dune sands From both transects but particularly from transect probably derived from a single sand source ofmixed 2, it was apparent that the upper loose sand was sediments. Since there is also no indication from the coarser than the basal slightly sand. The sand of the grain size distribution of great differences in the dunes of traverse 2 ismore uniform than those from distance over which sands have been transported, it is traverse 1 and the reason for this is the probably unlikely that the dune sands were transported by of the traverse proximity bed in 1. In wind across the whole dunefield, a result that is all of traverse 2 the loose sand is also samples upper (a) supported by the findings ofWasson (1983 a). well fine moderately sorted, skewed medium sand, The difference between upper and lower dune sand while the lower is less (b) sample invariably slightly probably indicates the effect of prolonged illuviation well sorted fine skewed fine sand (Fig. 3). moving shightly finer sand fractions from the upper In the from traverse 1 the same trend is samples sand into the lower horizon (see e.g. Corbett 1969, present but less well with the loose Twidale no defined, upper 1981).While there is difference in sorting sand being slightly better sorted and slightly coarser from one end of the dunefield to the other there is a than the lower sand. The sand is in upper moderately striking decrease feldspar content with increasing sorted coarse skewed to fine sand. symmetrical distance from the present lakes. Again this trend is 41/1987 46_Erdkunde_Band

Feldspar 24n % ?- .# . 20

18

16

14

12 0 o Samples traverse1 10" 0 Samples traverse2 8- o ?l o o ? . 6- ^ o 4- I # 2- . ? O-l-,-,-.-,-,-,-,-,-,-,-.-T-,-,-,-,-,-,-,-,-?-,-.-,-, 100 SO0 150 200 250km

across Fig. 4: Feldspar content in dunes Strzelecki Desert using results from Method 1

Feldspatgehalt (nach Methode 1) der Diinen entlang der Strzelecki-Wiiste

Table 1: Feldspar and total iron content of dune sands traverse 1. Table 2: Feldspar and total iron content of dune sands traverse 2 Note that Method 3 was not carried out on all samples. Feldspat- und Gesamteisengehalt der Dunensande von Dash indicates no analysis Traverse 2 von Feldspat- und Gesamteisengehalt der Diinensande Traverse 1. Beachte: Methode 3 wurde nicht fiir alle Sample Distance from Method 1 Method 2 Total Lake Frome % feldspar % iron Proben durchgefuhrt; dies wird durch einen Querstrich feldspar (-) vermerkt 1A 140 0.8 1.6 0.63 IB Sample Distance Method 1 Method 2 Method 3 Total 1.9 1.4 0.98 fromLake % feldspar % iron 2 A 130 0.9 0.4 0.77 Callabonna feldspar grains feldspar 3.32B 1.2 0.97 in 500 3 A 120 0.9 1.7 0.70 IA 210 0.6 6 2 0.35 2.03B 2.9 0.76 IB 1.6 23 0.69 4A 110 1.5 1.5 0.70 2 A 200 1.7 3 3 0.49 3.24B 1.1 0.76 2B 2.45 4 2 0.74 5 A 100 0.9 2.6 0.81 - 3 150 4.9 17 0.83 2.45B 3.6 0.97 4 A 105 4.6 4 - 0.40 90 6A 0.5 4.0 0.73 - 4B 6.2 4 0.43 2.86B 0.8 0.73 5 A 100 3.8 4 9 0.87 80 7A 1.4 1.5 0.67 5B 4.5 0 9.7 0.66 2.87B 1.7 0.91 6 95 5.5 13 - 0.40 70 8A 4.3 3.8 0.73 - 7A 50 6.2 6 0.46 7.58B 2.0 0.80 7B 8.5 15 - 0.83 60 9A 6.0 7.1 0.76 8 A 47 4.14 14 3.5 0.50 6.89B 7.2 0.84 8B 7.4 18 9 0.71 10A 50 9.5 8.3 0.64 9 A 35 6.9 4 - 0.44 108.4 B 4.2 0.84 9B 6.1 9 - 0.87 11A 40 15.5 11.1 0.87 10A 30 6.0 14 - 0.67 11 16.2B 10.4 0.85 - 10B10.8 15 2.13 12A 30 19.5 11.7 0.76 11A 25 9.3 13 9.8 0.12 12 13.2B 13.2 1.07 11 B 7.9 12 12.8 0.87 13A 20 21.7 15.5 0.81 - 12 24 10.3 15 0.61 13 20.3B 13.0 1.11 13 10 9.0 14 - 1.07 14 10 A 21.7 15.6 1.04 14 0 9.6 20 - 0.77 1422.8 B 16.5 1.20 Ernst Loffler andMarjorie E. Sullivan: The development of theStrzelecki Desert dunefields, Central Australia 47 more more pronounced and consistent in traverse 2 consists of Tertiary clays. Lateral transport of sand than in 1 (Fig. 4, Tables 1, 2), where the amount of must be considered in any hypothesis that tries to ex feldspar decreases from 1-2% at the outer margin of plain the Strzelecki Dunefields. Wasson (1983 a) has the dunefield to 15-20% near the present salt lakes. argued against this concept of longitudinal dunes de veloping from lake fringingdunes on the grounds that most of the sediments underlying the northern Strze Discussion and conclusions lecki and Simpson deserts are alluvial in origin. Ho wever he does agree that the aligned pans must be Both themechanical analyses and the iron analyses remnants of lacustrine strandlines of some kind ifonly indicate the likelihood that there was a relatively for the reason that "there is no other explanation for homogeneous sand source for the dunefield of the it". He also presents evidence from an area east of Strzelecki Desert and that the difference between the northern Lake Frome where in fact longitudinal du loose top sand layer and the often indurated "core" nes have developed from transverse lake fringing to of the dune may be due local reworking and dune ridges. illuviation of clay from top to base. There is no trend We do not see any real contradiction between Was in the amount of ironwith increasing distance from son's explanation that "the longitudinal dunes east of nor the present lakes, does the sand sorting show Strzelecki Creek have been formed by aeolian redi any significant trends except that the sand appears stribution of lacustrine sediment, lake-shore trans near slightly finer the present lakes. The slight in verse dunes and beach deposits, and fluviatile depo crease in Fe downwards in the dune profile indicates sits" (Wasson 1983 a, p. 112) and our Lake Dieri eluviation of clayminerals and sand coatings from the concept. Wasson*s (1983 a) findings that in the area upper loose sands and illuviation into the lower parts west of Strzelecki Creek the dune material derived causes of the dune. This eluviation/illuviation also frommuddy alluvium and in the southern Simpson the differentiation of the "dune profile" into slightly Desert from deflation of pans are also not contra coarser well sorted upper sands and slightly finer less dicting our concept since there isno denying that ulti well sorted lower sands. mately all the dune material derived from the uplands The most factwith to our Lake to the northeast and north was important respect east, and brought into Dieri concept, however, is the clear trend towards an the basin by major river systems draining the eastern increase in the quartz/feldspar ratio with increasing highlands. The question, however, as to how these distance from the present lakes and hence an increase deposits were distributed across the dunefields trans inweathering and relative age across the dunefield. verse to the general wind direction can as far as the while we do not know how the lake receded and how Strzelecki Desert is concerned not be answered by al significant fluctuations and occasional regressions luvial redistribution alone. Lateral transport of the can no may have been there be doubt that the outer sediments along a former lake shore must be in margins of the lake were exposed firstand that the volved. dunes this area area occupying and the downwind of Our original interpretation represented an attempt it were was formed from the sand which exposed to explain a pattern clearly visible on the satellite ima earlier while the dunes furtherwest were and south gery and the striking pattern of aligned pans is still formed from sand later. one our exposed of strongest arguments for the existence of The littoraldunes of Lake Dieri probably provided Lake Dieri. From the satellite view coming down a uniform sand source of single mixed sediments to earth has proved more difficult and unequivocal derived from drainage into the lake basin from pre evidence for the existence of the lake has so far not and dominantly quartzitic granitic rocks exposed been found. The published stratigraphy of the lake in the to the N and E. This as a uplands function basin is complex (cf.Wasson 1983 a) and there ist still distributor of sand across the no on laterally lake and along information the timing of the lake or the events the lake shore hence across (and the present dunefield following its disappearance (Callen 1977, King at to the wind right angle predominant direction) 1956,Wopfner a. Twidale 1967). However we feel from themouths of themain rivers is an important that the data presented here and in particular the role as the sand of the Strzelecki Desert can not clear trend of thequartz/feldspar ratiowith increasing have derived from the small of catchments the rivers distance from the present remnant lakes do support draining the Flinders Ranges nor can ithave derived our view of the existence of a formerLake Dieri follo from a local source beneath the present dunefield wed by the establishment of the present dunefields in which in the central and southern Strzelecki Desert the Strzelecki Desert. 48 Erdkunde Band 41/1987

sands. In: in Acknowledgements Progress Physical Geography 5, 1981, 514-534. D.: The record at Lake The work on which this paper is based was under King, Quaternary stratigraphie takenwhile the authors were members of theDivision Eyre North and the evolution of existing topographic forms. In: Trans.R.Soc.S.Aust. 79, 1956, 93-103. of Land Use Research (now Division ofWater and Loffler, E. a. Sullivan, M. E.: Lake Dieri resurrected: Land Australia. The Resources), CSIRO, laboratory an interpretation using satellite imagery. In: Z. Geo were carried out theAustralian Mineral analyses by morph. N.F. 23, 1979, 233-242. Laboratories and financed the Development by Twidale, C. R.: Age and origin of longitudinal dunes in Land is Division of Use Research. This gratefully the Simpson and other sand rigde deserts. In: Die Erde acknowledged. 112, 1981, 231-247. Wasson, R. J.: The Cainozoic history of the Strzelecki and Simpson dune fields (Australia), and the origin of the de sert dunes. In: Z. Geomorph. N.F., Suppl. Bd. 45, 1983 (a), 85-116. - : References Dune types, sand colour, sediment provenance and hy drology of the Strzelecki-Simpson dunefield, Australia. a. Callen, R. A.: Late Cainozoik environments of part of In: Brookfield, M. E. Ahlbrandt, T. S. (Eds.): Eo northeastern Southaustralia. In: J. Geol. Soc. Aust. 24, lian Sediments and Process, Amsterdam 1983 (b), 1977, 151-169. 165-195. H. a. Corbett, J. R.: The Living Soil. Sydney 1969. Wopfner, Twidale, C. R.: Geomorphological Folk, R. L.: The petrology of sedimentary rocks. Austin, evolution of the , central Australia. In: a. Texas 1974. Jennings, J. N. Mabbutt, J. A. (Eds.): Landform stu Gardner, R. a. Pye, K.: Nature, origin and palaeo dies from Australia and New Guinea. Canberra 1967, environmental significance of red coastal and desert dune 118-143.

BERICHTE UND MITTEILUNGEN

URBAN HEAT ISLAND DEVELOPMENT IN MEDIUM AND LARGE URBAN AREAS IN MEXICO

With 4 figuresand 1 table

Ernesto Jauregui

1. Introduction zone that it seems possible now to draw some general izations regarding itsmorphology and time variations on for Oke A typical example ofman-made modifications (see example 1982). urban climate studies in the are the city climate due to the urbanization process is the By contrast, tropics so-called heat island effect. The phenomenon of air relatively recent and very sparse. The few studies that in cities have been temperature often being higher in the city than in its have been undertaken tropical on thermal contrasts estab surrounding country side, has long been recognized mainly based urban/rural and documented formid-latitude cities. Studies on lished from climatological records. The need for heat island effects are so numerous for the temperate information on tropical urban climates has been