JOURNAL OF GEOPHYSICAL RESEARCH, VOI. 101, NO RI1, PAGES 27,503-27,516, NOVEMBER 10, 1998 Emplacement of long lava flows on planetary surfaces James R. Zimbelman Center for Fanh and Planetary Studlcs, Nattonal Arr and Space Museum, Srnlthsontan Initttut~on, Wailungton, D C Abstract. Three long lava flows on Mars, Venus, and the Moon were examined in order to evaluate their possible emplacement rate and condition. On the Moon, flows o f the lay1 (phase 111) effusion within the Imbrium impact basin were examined using Apollo photography. The longest phase 111 flow can be followed for 250 km, terminating -400 km from the probable source vent. This flow has a width of 10 t o 25 km, thickness of 10 to 30 m, and a medial channel preserved in its proximal reach, and it was emplaced on a regional slope of -0.3". In the Tharsis region of Mars, a well-defined set of lava flows extends north from the topographic saddle between Ascraeus and Pavunis Montes. Viking Orbiter images show one flow h a t can b e triaced Tor 480 km, with a width ranging from 5 to 50 km, thickness of 3 0 to 100 m, and a prominent medial channel in its proximal reach, and was emplaced on a regional slope of -0.5' to -O.lO. The Strenia Fluctus area on Venus consists of an m a y of intermixed radar-bright and radar-dark lobate flows, one of which can he traced for 180 km, with a width of 5 to 20 km, and an unknown thickness (hut inferred to be -30 m), and was emplaced on the lowland plains where the regional slope is only -0.03'. When viewed at the full Magellan resolution, this flow contains several flow margins, indicating its compound nature. Effusion rates were calculated for the simple lunar and Martian flows using published empirical and theoretical relationships, resulting in a broad range of 500 to 108 (Moon) and 600 to 2 x 108 (Mars) m3/s, with most likely values of -5 X lU4 t o -lo5 for both flows. The compound Venus flow would have required 494 years for emplacement at the typical Kilauea rate of -5 m3/s, but the thermal balance of planetary tube systems could also be consistent with a rate at least an order of magnitude larger. The distinction between simple and compound flows is important to any evaluation of flow emplacement based solely on remote sensing data. 1. Introduction Volcanic flows of great length have becn observed on scveral planetary surfaces [e.g., Lopes-Goutier, 19931. Long lava flows on three planetary surfaces are examined here to provide constraints on the emplacement conditions experienced by these lavas. The observable flow dimensions, the topography over which the lava flowed, and the environmental conditions are compared for long lava flows on the Moon, Mars, and Venus. The results should provide a basis for comparison with long lava flows on Earth and other planetary surfaces. The three lava flow localities chosen for this study represent distinct settings in which long lava flows have been documented with spacecraft images and photographs. Each lung flow has traceable flow margins in excess of 180 km in total length, and each occurs within broader flow fields indicative of substantial local volcanic cffusion. The specific sources for the three flows are not identifiable, so the observed flow lengths are minimum values only. All three flow locations occur in areas of broad volcanic plains with relatively low relief. 2. Background Lung basaltic lava flows on Eanh are usually associated with localized massive volcanic effusion in areas called "large Copyright 1998 by thc American Geophysical Union. igneous provinces" (LIPS). LIPS are widely distributed o n Earth, ranging i n size from the enormous Ontong Java region on the east Pacific seafloor (-3.6 X lo7 km3) to the smaller hut extensively studied Columbia River Basalt (CRB) group (-1.3 X 1u6 km3) in the Pacific norfhwest of North America [Coffin aid Eldholm, 19931. LIPS occur as oceanic plateaus (e.g., Ontong Java), on volcanic passive margins (un continental shelves), or within continental land masses (e.g., CRB). These diverse settings lcad to a wide range in the state of knowledge of each province due to the difficulty or ease with which the rocks can be accessed, but still an impressive global data set has resulted from individual studies [Mucdougull, 19881. Recently. a debate has arisen concerning the mode of lava emplacement within large igneous provinces, specifically within the CRB continental flood basalts. Shaw Md Swanson [I9701 provided the first quantitative estimates of CRB emplacement, relating constant effusion along fccdcr dikcs to turbulently flowing lava powered by the hydraulic head generated along the gently sloping lava surface. Their calculations resulted in emplacement times of days to a few weeks for CRB flows, encompassing individual flow units hundreds of kilometers in length, driven in large part by the observation that very little chilled glass is preserved in CRB flows, which was interpreted to imply extremely rapid emplacement [Shaw and Swanson, 19701. This approach was adopted by subsequent studies of the emplacement uf long lava flows on the Moon [Srhnb~r, 1973hl and Venus [Roherfs et al., 19921. However, recent observations of basaltic lava flow inflation in Hawaii, caused hy the rise of an initially emplaced surface through continued influx of lava, have provided a 27,504 ZIMBELMAN: LONG LAVA FLOWS ON PLANETARY SURFACES mechanisn~ lor the slow emplacement of what eventually becomes a thick lava flow [Walker, 1991; Hon ef al . . 19941. Field evidence cited for inflation features within the CRB [Self er al., 1996, 19971 remains controversial, sincc many of these areas are also interpreted to be consistent with the rapid emplacement of the lava [Swanson er al . , 1989, pp. 21-26; Tolan el al. , 1989; Reidel and Tolan, 19921. If the lavas were emplaced via inflation, flow units within the CRB could require years to decades for emplacement, in marked contrast to the days to weeks envisioned by the flood scenario. Hopefully, insights from long planetary flow studies can help clarify distinctions between scenarios and point to helpful evidence from other studies of long lava flows. The issue of fast versus siow emplacement of the CRB flows was addressed during a field trip preceding the 1995 International Union of Geology and Geophysics (RIGG) meeting [Hun and Palliuler, 19951, where participants were shown some of the controversial features while advocates for both fast and slow emplacement presented their cases. Both sides made sufficiently compelling cases to warrant further study, but both sides require additional field cvidcnce addressing both fast and siow emplaccmcnt within the flows of the CRB flows. A Chapman Conference in 1996 specilically addressed the conditions of emplacement for long lava flows, and an associated field trip visited localities on the Undara, Kinrara, and Toomba flows, Cenozoic basalts that include some of the longest subaerial lava flows in the world [Srephenrun ond Whirehead, 19961. Once again the issue o l fast versus slow emplacement was integral to the field localities, particularly for assessing the field evidence for inflation within the long lava flows. The Australian flows displayed pahoehoe textures over essentially their entire lengths, one of which is 160 krrr lung [Stephenson el al . , 19961. Lava tubes are important in discussing possible emplacement conditions for long lava flows. Several descriptions are present within the literature for the formation ul lava tubes [e.g., Greeley, 1971, 1987; Perer~on ruzd Swanson, 1974; Hon el al . , 1994; Peterson er al . , 19941. Recent studies of the thermal effect of transport of lava through a tube illustrate the dramatic effect the insulating properties of a well-developed tube can have on the lava carried by the tube [Kesrrlzelyi, 1995; Dragoni er al., 1995; Sakin~oto er al. , 19971. In particular, tubes provide a viable mechanism for transporting basaltic lava for hundreds of kilometers within the environmental conditions encountered on all of the terrestrial planets [Keszrh~lyi , 19951. Tube-fed levas are inferred to have played a prominent role in the emplacement of thr. 120-km-lung Tuomba flow in North Qucer~sland [Whirehead and Srephenson, this issue] and the 75-km-long Carrizozo flow in New Mexico [Kesztheyli and Pieri, 19931, as well as for many flows on thc occan floor [Ballard er al., 1979; Bonalri ond Harrison, 1988; Gregg and Fink, 19951. Two ways lava tubes form in basaltic flows are roofing over of active lava channels [Peterson er al. , 1994; Dragoni er al . , 19951 and locally concentrated flow inside an inflating flow lobe [Hon er al., 19941. On planetary flows a roofed-over lava channel might be inferred from a medial (leveed) channel and aligned collapse pits. Aligned collapse pits have been cited as evidence for tubc-fkd ernplacernrnl on surrle planetary lava flows [Oberbeck el al. , 1969; Chrr er al. , 1977; Mouginis- Mark er ol. , 1988; Sakimoto et al . , 19971, but none of the flows described below display this distinctive pattern. The relationship between lava tubes and inflatcd flows is clearly relevant to a discussion of possible effusion rates associated with planetary flows, but it is unlikely that features diagnostic of inflation can be detected unambiguously using the planetary remote sensing data currently available. The transition from thc smooth, ropy texture of pahoehoe lava to the clinkery rubble of an aa flow is clearly related to effusion rate for Hawaiian flows [Rowland and Walker, 19901. However, this transition is also closely related to the rate of shear experienced by the lava and its volatile content [Kilburn, 1990, 19931. The presence of pahoehoe or aa surface lextures is dilficult to deter~ninc using current planetary image data, although the use of varlable rod lengths in measuring the margins of lava flows shows great promise for distinguishing between the fractal nature of these flow types [Bruno el a/.. 1992, 19941. Flow texture is strongly influenced by the environnlent into which the flow is emplaced, along with the effusion rateof the flow [Griffiirhs and Fink, 1992a. b; Gregg and Fink, 1995, 19961, so the pahoehoe-aa transition will not be used here to infer effusion conditions. A more fundamental issue is the distinction described by Walker [I9711 as "compound" or "simple" lava flows. Compound flows consist of multiple flow fronts and margins emplaced as the flow advances by near-simultaneous effusion from many locations along the front, often highly intermixed within the broader confines of the flow outline. By contrast, simple flows are emplaced behind a single advancing front, and they can d~splay a medial channel which supplird the active snout during emplacement if conditions allowed the channel to drain once effusion stoppcd. The distinction bctwccn compound and simple flows can provide a clue to the flow emplacement, and some of these characteristics r e resolvable with existing planetary data. 3. Three Planetary Examples of Long Lava Flows The lava flows examined here are found on broad volcanic plains on the Moon, Mars, and Venus. The associated flow fields occur in vcry diffcrcnt physical environments, ranging from a dense carbon dioxide atmosphere (Venus) to a vacuum (the Moon), and yet they all display comparable overall dimensions and plan views. In this section we describe the characteristics of a long flow within each field, as well as the general aspect of the field in which it occurs. Basic parameters for the long flow identified on each plane1 are summarized in Table 1 3.1 Mare Imbrium, Moon The large impact basins on the Earth-facing side of the Moon are flooded with basaltic lava that was sampled by five of the six Apollo landings [Vanimon er a / , 19911. Mare Imbrium includcs some of the last lavas crupted onto the basin floor, with many clear flow margins and occasional levecd channels on flows that are traceable for hundreds of kilometers ISchaber, 1973a, b; Schaber et al. , 19761. The early phases of the Imhrinm emptions generated lava flows that extend 1200 and 800 km (for phases I and 11, respectively) from the vent area [Schaber, 1973a, b]. Thelast eruptive episode (phase Ill) generated flows that traveled 400 km (Figure 1) and are traceable close to the inferred source vent near the Euler-8 feature located at 2Z050'N. 3I020'W [Schaber. 1973al. ZIMBELMAN: LONG LAVA FLOWS ON PLANETARY SURFACES Table 1. Characteristics of Long Lava Flows on Three Planetary Surfaces p~ Region Mare lmbrium Tharsis Montes Strenia Pluctus Planet Figure Location Latitude, "N Longitude, " Dimensions Lcngth,km Width, km LengthnVidth Thickness, m Area, km2 VolumeS h3 Average slope, Environment Surface temperature, K Atmosphere Atmospheric pressureb, bars Surface gravity, m/s2 ~~~ Moon I 26 to 32.5 21 1028 W 250 I0 to 25 10:l-25:l 10 to 30 4800 48- 144 0.3 -1Wto-360 vacuum 0 1 62 Venus 6, flow 1 -150 to -290 carbon dioxide -750 c h o n dioxide ' h a multlphed by ran e of thicknesses % bOne bar pressure = LO dynlcm2 = LOS N/m2 I I I I I 0 Heis - 0 Carlini B La Hire - 100 km - I I I I I I 3zDW 28OW 24OW 20? W Figure 1. Flow margins of phase 111 lava flows in Mare Imbnum on Earth's Moon latter Schaber, 1Y73a. Figures 30-15 and 30-21bl. Large Impact craters and hills (pattern) are shown for reference. The flows progresses to the north-northwest from a vent area inferred to be along the southeastern margin of the Euler-B feature at 22"50'N, 31?20'W [Schaber, 1973hl. The flows are interpreted to be compound within 120 km of the source vent [Schaber et al., 19761. Shaded flow represents the longest simple flow traceable beyond the compound flow region: see Table 1 for dimensions. 27,506 ZIMBELMAN. LONG LAVA FZOWS O N PLANETARY SURFACES Figure 2. The phase 111 lava flows i n Mare imhrium photographed at very low solar incidence angle (ZO) during the Apollo 15 mission. The flows traveled north-northwest. La Hire mountain is at right center. A portion of the longest traceable flow (shaded in Figure 1) passes through the center of the frame, including a leveed medial channel (arrow) representing about one half of the flow width in this area. Photogrammetric analysis in Schaber- el al. 119761 indicate flow margin heights in this region rangc from 7 to 36 m, rcduced from the 12 to 53 m reported by Schrrbrr [1973b]. Portion of Apollo 15 metric camera photograph AS15-M3- 1557. Excellent low-illumination Apollo photography provides good documentation of ihe morphology of these latestage lunar mare flows. Flow margins and medial (leveed) channels (Figure 2) were recorded in photographs ohtained at solar illumination angles of from 4' to ZO, which highlight these ~ u b t l e features [Sckabcr, 1973a, b]. The ubiquitous impdct cralers un the 1mbriu111 flows indicate that they have been subjected to a prolonged history of cosmic bombardment. No Apollo samples were obtained from this area of the Moon, hut the cratering record suggests that these flows are among the youngcst materials emplaced onto the lunar surface, with probable exposure ages of from 2.5 to 3 . 0 b.y. [Schaber el a l . , 19761. The lmbrium lava flows show distinct differences from other mare materials at infrared and ultraviolet wavelengths [Wltitaker, 19721, consistent with color ratio images of relatively young terrains elsewhere on thc Moon. Laboratory analyscs of lunar-like matcrials have documcntcd lhc rclativcly low viscosily (5-9 Pa s at 1250-C [Weill er al., 19711; 1 Pa s = 10 poise) and high bulk dcnsity (2950 kglm3 [Murase and Mctlirney, 19701) of lunar basalts, as compared to typical terrestrial basalts. The Apollo samples provide the only rheological properties available at present that are directly applicable to lava flows observed on planetary surfaces, as well as the only case in which we are certain that the specific planetary lavas are basaltic in nature. Observed flow widths for the phase 111 lavas range from 5 to 25 km fnr flows >I50 km from the inferred vent. The phase 111 flows attain a combined width of -120 km at a distance of -120 km from the vent area (Figure I). Sclraber er al. [I9761 ~nfer that the proximal phase Ill flows were emplaced as compound, overlapping flows. Consequently, we will restrict our discussion to the longest "simple" phase 111 flow (shaded in Figure 1) that attained the greatest distance from the vent area of any of the phase 111 lavas. This flow can bc trnced for ZIMBEUIAN. LONG LAVA FLOWS ON P W A R Y SURFACES 27,507 250 km with a variable width of from 10 to 25 km, resulting in a lengthlwidth aspect ratio of from 10:l to 2 5 1 . A suhtlc medial channel is present on the proximal reach of this flow, corresponding to about one half o f the flow width in the channeled reach (Figure 2). Photogrammetric measurements of the phase 111 flow heights initially ranged from 10 to 63 m [Schaber, 1973b1, but later analysis reduced this range to 10 to 30 m [Schaber et al . , 19761. The lunar environment is notable in that it is now and likely has been for billions of years, a nearly perfect vacuum. Thus cooling of lunar lava flows has been by radiation and conduction, without the convection component available for flows emplaced within an atmosphere or on the ocean floor. Solar illumination generates surface temperatures from -100 to -360 K throughout the lunar day, equal to its 29-day orbital period around Earth. The relatively small size of the Moon results in a low value for the acceleration of gravity (1.62 m/sZ) as compared to that on Earth (9.80 m/sz). Topography within the Imbrium basin is very subdued, with an average slope from the source region to the basin center estimated to be between 1:100 (0.57O) and 1:1000 (0.06") [Schaber, 1973bl. The longest phase I11 flow can only be followed to -120 km from the vent area, where slopes are likely to be considerably less than the maximum value near the source. Recent laser altimetry from the Clementine spacecraft [Zuber er al., 19941 show that the distal phase I flows in the Imbrium basin are topographically higher than the basin center, suggesting that significant basin subsidence took place subsequent to the emplacement of the phase I and I1 Imbrium lava flows. 3.2 Thars i s Montes, M a r s The three large Martian shield volcanoes which compose the Tharsis Montes occur near the summit of the 4000-km-wide Tharsis uplift, where the Martian crust is elevated -10 km above the surrounding terrain [Carr, 1981, pp. 87-95], Each of the three shield volcanoes is approximately 200 km in basal diameter, with relief of 12 to 17 km above the surrounding summit of the Tharsis uplift [US. Geological Survey [U.S. Geological Survey (USGS), 19911. All three volcanoes are surrounded by volcanic plains composed of numerous flows interpreted to be derived from either the volcanoes themselves or nearby unidentified vents [Scon and Tanakn, 19861. A broad fan of lava flows southwest of Arsia Mons has been discussed at length in the literawre, with some flows up to 700 km i n length [Carr er al. , 1977; Moore et al. , 1978; Schaber ef a l . , 1978; Scott and Tanaka, 19811. Another well-defined set of lava flows emanates to the northwest from the topographic saddle between Ascraeus Mons and Pavonis Mons volcanoes [Scoff et al., 1981a, bl. Images from the Viking Orbiter spacecraft reveal one lava flow within this field whusr margins 5' N can be traced for 480 km, as well as several other adjacent 112Ow 108' flows that follow the same general trend (Figure 3). The vents Figur* 3. Flow margins visible northwest of the for these flows cannot be identified since subsequent lava flows topographic saddle between Ascraeus Mons and ~ a v o n i s ~ o n s have buried the proximal portions of all of thge visible flows. on Mars. Shaded flow has margins that can be traced for 4 8 0 km down the northwestern flank of the Tharsis uplift; see Table The 480-km-long lava flow (Figurc 3, shadcd) shows some for dimensions, important variations in basic morphology and dimensions over its traceable Length. The proximal portion of the flow i s gently sinuous with n medial channel, flanked by natural lava quite uniform at -5 km. The leveed medial channel implies that levees. that accounts for about one half of the flow width at least this portion of the flow likely was fed by a river of lava (Figure 4). Shadow measurements indicate that the proximal comparable to the channel-fed na flows common in Hawaii flow margins arc -30 m thickness. The proximal flow width i s [Lipman and Banks, 1987; Rowland and Walker, 19901. ~ ~ 27,508 ZIMBELMAN: lQNG LAVA FLOWS ON PLANETARY SURFACES Regional topography i n the proximal portion of the flow has an average gradient of -0.5" toward the northwest [USGS, 19911. The medial channel disappears once the flow reaches the more gently sloping (-0.1') plains surrounding the Tharsis volcanoes (below the 5 k m elevation contour, [USGS, 19911. The flow widens to more than 30 km on the gently sloping plains, locally attaining a width of 50 km. Distal portions of the flow display broad upper surfaces lacking any distinct morphology at the 190 mlpixel resolution of the best Viking images of this area (Figure 5). Shadow measurements of margins near the terminus of the flow indicate a thickness of -100m. Owing to the highly variable width along the flow, lengthlwidth aspect ratios range from 10:l for distal widths to 96:l for proximal widths. Numerous other lava flow margins are evident surrounding the long flow (Figures 3 and S), indicating that several episodes o f lava emplacement occurred in this area [Scoff er al., 1981a. b]. Many of these earlier eruptions must have been longer than the flow described above, which likely was one of the last in the eruptive sequence. Here we assume these flows are approximately basaltic in composition in the absence of compelling compositional evidence. The precise age of the Martian flow cannot be determined, Figure 4 ' portion of long lava "Ow but the flows surrounding the Tharsis Montes are from the early (arrows) in Ihe Montes legion Of Mars' A leveed Amazonian to late Hesperian eras which are relatively late in medial channel representing about one half of the flow width i s evident along the proximal reach, Portion of Viking Orbiter the stratigraphic sequence of terrain units identified on Mars image 643A51, orthographic projection, neuttal gain filtered [Sconund Tanaka, 19861. The Martian atmosphere likely has version, 270 mlpixel resolution, centered on 7.I0N, 109.8"W. not varied greatly in overall composition during the last half Figure 5 . Distal portion of traceable long lava flow (arrows) in the Tharsis Montes region of Mars. Numerous other flow margins are also visible in this region, although the individual flows cannot be traced for lengths beyond -100 km (Figure 3). Shadow measurements along the traceable long flow indicate flow Ihicknrss is -100 m in the distal reach. Portion of Viking Orbiter image 516A52, orthographic projection, neutral gain-filtered version, 190 mlpixel, centered on 13.0?N, 109.1?W. ZIMBELMAN. LONG LAVA FL.OWS ON PUNETARY SURFACES 27,509 Fieure 6. Flow mareins visible in the Strenia Fluctus region of Kawelu Planitia on Venus. The intermixed . . tlows h v c SAR rzilc~t.inrc~s that arc mtlar-hriphl ;,"(I rd.l;ir-J~rk .,cc F I ~ U I C 7 S11~de.l ;mJ ~ ~ u n ~ h e r e d l l ~ ~ w \ J r i dlscusrcd in the rev: srr T:~hlc I lor d~mrnsioni of' ll81u I P:ttlsrn ~ n J ~ c x c \ <)ut.r.>ps ,f ~ntenrel! Jcfunncd rock (tessera) that predate the lava flows. Dots indicate individual low domes, one of which was the source of the shaded flow field 4. of Martian history, but orbital variations could have resulted in considerable variability in total atmospheric pressure [Pollack. 19791. Given the uncertainty of the precise time of the flow eruption, we will take the current atmospheric conditions to be representative of the Martian atmosphere during flow emplaccmcnt. The location of the flow field on the flank of the Tharsis uplift means that the current atmospheric pressure is reduced to ahnut -4 mhar (610 ~ l m ~ ) . Thermal infrared lneasurernents fro111 the Viking IRTM experiment indicate that surface tcmpcrawrcs in the flow area can range diurnally from -150 K to -290 K [Kieffer a1 a / . , 1977; Zimbrlman and Kieffer, 19791. 3 . 3 St ren ia Fluctus, Venus The Magellan mission provided the first detailed views of nearly the entire surface of Venus through the use of synthetic aperture radar (SAR) imaging [Saunders et a / . , 19921. The lowland plains of Vcnus contain abundant volcanic flow features [Head er a [ . , 19921. Several studies have focused o n various aspects of the volcanic features on Venus, hut the one most relevant to the current study was an analysis of the Mylitta Fluctus flow field [Roberts et al., 19921. Components of the Mylitta Fluctus flow fleld extend up to 1000 km from the inferred source region, representing one of the areally most extensive volcanic effusions documented on the planet [Head et a / . , 1992; Roberts er al . , 19921. Here we focus on a lava flow field more typical of those that occur throughout the Vcnusian lowlands. The Kawelu Planitia region includes numerous volcanic flows, most of which emanate from concentrations of small volcanic domes [Zimbelman, 19941. Recently one flow field in Kawelu Planitia was given the provisional name of Strenia Hnctus (Figure 6) . The Strcnia Fluctus flow margins lose their identity approaching the source domes (off the left margin of Figure 61, but individual flow components can be traced for r~early 200 km. The flows cunaisl o r intern~ixed radar-bright and radar-dark lobatc units, where the SAR imngc brightness i s a complex interaction of the radar signal with surface roughness elements and the dielectric properties of those surfaces (Figure 7). Several radar-bright flow segments can be traced for >I00 km, with widths that range from 5 to 20 km. When viewed at full resolution, each of these radar-bright flows are seen to consist of numerous superposed margins indicative of a compound flow (inset, Figure 7). A prominent -150-km-long radar-dark flow (left center, Rgure 7) is -16 km wide at its proximal (upslope) reach but it narrows to -2 km width down slope, where it flowed between previously emplaced flows; this flow is comparable in planform and scale to the Toomba flow in North Queensland, Australia [Stephenson and Whitehead, 19961. Both the radar-bright compound flows and the radar-dark flows in Strenia Fluctus are consistent with quantified studies of radar backscattcr in the Magellan images that indicate most flow features on Venus have surfiace roughness comparable to pahoehoe flows o n Earth; only the very brightest flows on Vcnus have roughness approaching that of terrestrial aa [Campbell mui Campbell. 27,510 ZIMBELMAN: LQNG LAVA FLOWS ON PLANETARY SURFACES Figure 7. Strenia Fluctus lava tlows on the lowland plalns of Kawelu Planitia, Venus (compare with Figure 6). Superimposed contours are elevations in meters above 6250 km radius, derived from Magellan radar altimetry measurements. The radar-bright flows at center (flows 1 to 3 in Figure 6) consist of numerous overlapping flow lobes when viewed at full resolution and optimized contrast (see inset. where letters indicate relative superposition with " d o n top and "a" on bottom). Black arrow indicates a 2-km dome that is the solrrce fnr a flow complex (flow 4 in Figure 6) that traversed a regional slope of 0.03'. A radar-dark flow at the left center starts with a width of -16 km but narrowst0 a width of -2 km between previously emplaced flows; this flow is comparable in planform and length to the Taomba flow in Nonh Queenslwd, Australia. Portion of Magellan Cl-MlDR 45N244, tiles 46. 47, and 48. Inset is from portions of F-MIDR 40N251, t ~ l e s 21, 2 2 , 29, and 30. 19921. Here we again assume an approximately basaltic composition in the absence of composition data. For comparison with long flows on other planets, wc focus here on the radar-bright compound flows in Strenia Fluctus (flows 1 lo 3 in Figurc 6). Thrcc bright flows extend southeast past a n outcrop of faulted tessera terrain, and their sources are buried beneath subsequent flows. All three tlows are compound in nature when examined at full resolution and optimized contrast. These flows extend down a regional slope of 0.05' to 0.03O (contours, Figure 71, which corresponds to a vertical drup of lo: I ) that are difficult Lo recuricilc with rapid eCiusion on a very shallow slopc [Zimbelman, 19961. In particular, it seems difficult for a very large extrusion rate, whether from a point source or a linear dike, to produce a flow many times longer than it is wide, especially on a regional slope of -0.1'. Recently. Miyamoto and Sasaki [this issue] carried out numerical simulations of lava flow emplacemenr which showed that flow width appears to he sensitive to effusion rate for large simple flows; i n general, wider flows required a higher effusion rate than narrow flows. From this perspective, the hroad width of the proximal phase 111 lmbtium lavas (Figure 1) may be consistent with relatively large initial effusion similar to that modeled b y Shaw and Swanson [1970], but no definitive evidence of turbulent flow has yet been demonstrated for terrestrial lava flows. The relatively narrow planform of the distal phase 111 lmbrium lava and thc othcr planctary flows considered here are not easily accommodated by flood-like effusion. The interaction of tlow segments with subtle topographic features, such as i n the Strenia Fluctus flows o n Venus (Figure 7). also suggest the flows were ahle to respond to changes in slope of the order of only 0.03.', favoring slow rather than fast effusion. While it is not possible at present to state conclusively that such considerations require modest Hawaii- like effusion rates, it i s difficult to reconcile the three planetary flows examined here with the massive effusion rates and turbulent conditions attributed to flood-like emplacement. 6. Summary Primary conclusions are as follows: (1) The long flows o n the Moon and Mars are interpreted here to be simple flows from the lack of multiple flow margins and the presence of a medial channel in their proximal reaches. This observation allows their effusion rate to be estimated using several published relationships involving flow dimensions. ( 2 ) The lunar and Martian flows have a wide range of calculated effusion rates, depending on which relationship is used, but the most likely values are -5 x lo4 to -lo5 m31s for both flows. (3) The Venus flow displays intermixed flow margins within the flow outline, suggesting it i s compound. If erupted at rates typical of tube-fed eruptions on Hawaii, this flow required 494 years. (4) It is helpful to evaluate whether flows are simple o r complex prior to estimating effusion rate, particularly when the flows can only b e examined with remote sensing data. Acknowledgements. The comments of reviewen Tracy Grcgg, Christopher Kilbum, and Associate Editor P.J. Stephenson were extremely helpful in cluifying results presented in exly versions of the manuscript. This work was suppatted by NASA grants NAGW-3364, NAGW-5000, and NAGS-4164 The help of A. Johnston in the digitization and preparation of Figures 1, 3, and 6 is greatly appreciated~ References Ballard, R. D., R.T. Holcomb, and T. H. van Andel, T.H., The Galapagos rift at 86'W, 3, Sheet flows, collapse pits and lava lakes of the rift valley, J . Geophys. Rex., 84, 5407-5422, 1979. Borralti, E.. and C. G. A. Harrison, Eruption styles uf basalt in occanic sorevdine ridges and seamounts. Effect of )maenla tenlocraturc and viscosity, J. Geophys. Res., 93, 2967-2980, 198;. Bruno, B. C.. G. J. Taylor, S. K. Rowland, P. G. Lucey,and S. Sclf, Lava flows an: iractiib, Geophys. Rcs. Len., 19(3), 305-308, 1992. Bruno, B. C.,G. J. Taylor, S. K. Rowland, and S. M. Bnlogn, Quantifying the effect of rheology on lnvn~flow mugins using fmtnl geometry, Bull. Vnlcanol., 56. 193~206, 1994. Cnmpbell, B. A., and D. B. Campbell, Analysis of volcanic surfacc morphology on Venus from comparison of Arecibo, Magellan, and terrestrial airborne radar data, J. Geoph3is. Res., 97, 16293~16314, 1997 C m , M. H., Geologic map and section of the Timochans region of the Moon. U. S. Geol. Surv Map. 1-462 (LAC-40). scale 1.1,000.000, 1965. C m , M. H., The Surfoce ofMars. 232 pp., Yale Univ. Press, New Haven, Conn.. 1981. C m , M. H., R. Greeley, K. R. Blasius, J. E. Gucst. and 1. B. Murray, Some Martian volcanic features as viewed from the Viking orbiters, J. Ceophys. Res., 82, 3985-4015, 1977. Coffin, M. F.. and 0. Eldhom, Large igneous provinces, Sci. Am., 269(3), 42-49, 1993. Crisp, J., and S. M. Baloga, Influence of crystallization and entrainment of cooler material on the emplacement of basaltic an lava flows, J. Ceophyr. Res.,99, 11819-11831, 1994.