The role of rifting in the generation of melt: Implications for the origin and evolution of the Lada Terra-Lavinia Planitia region of Venus

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL 100, NO El, PAGES , JANUARY 25, 1995 The role of rifting in the generation of melt: Implications for the origin and evolution of the Lada Terra-Lavinia Planitia region of Venus Karl P Magee and James W Head Department of Geological Sciences, Brown University, Providence, Rhode Island Abstract A rift zone over 6000 km in total length runs along the border of Lada Terra, a highland in the southern hemisphere of Venus, and Lavinia Planitia, a basin that has been interpreted as a site of early-stage mantle downwelling Along the length of the rift are a number of volcanic centers of widely varying morphology and volcanic output These include coronae, radially fractured domes, and large flow fields similar in scale to terrestrial flood basalts We develop a model for the origin of extension related to passive rifting in response to stresses created by the adjacent downwelling Volcanism and extension at other rifts on Venus, such as Devana Chasma, have been attributed to deep-seated mantle plume activity In contrast, we interprethe origin of extension and volcanism along the Lada rift to be linked to upwelling and decompression melting of mantle material due to rifting and, possibly, to counterflow associated with downwelling Extension occurred generally prior to the formation of volcanic centers and the eruption of large-scale 11ow fields, although most of the volcanic centers have been fractured by continued extension along the rift Current debate over the formation of terrestrial flood basalts centers on the necessity of preexisting extension and stretched and thinned lithosphere to produce enhance decompression melting within a large plume head or mantle thermal anomaly Our studies of large-scale flow fields associated with the Lada rift and coronae on Venus indicate that extension is a prerequisite for the formation of the majority of large-scale flow units on Venus Introduction importance of lithospheric extension in the generation of melt from mantle thermal anomalies [eg, McKenzie am:l Bickle, In the southern hemisphere of Venus a rift zone over ; Richards et al, 1989; White and McKenzie, 1989; km in length runs along the eastern and southern boundary of Campbell and Griffiths, 1990; Hooper, 1990; Hill, 1991; Lavinia Planitia, a lowland plains region, and Lada Terra, a Anderson et al, 1992; Saunders et al, 1992; White, 1992] southern highland [Campbell et al, 1991; Senske et al, 1991; Roberts et al, 1992] Along this rift are multiple volcanic Regional Setting centers Included are coronae, features thoughto be the surface Lavinia Planitia is a bowl-shaped basin centered at manifestations of mantle upwellings impinging against the lithosphere [Pronin and Stofan, 1990; Stqfan and Head, 1990; approximately 45øS, 345øE (Figure 1) It has a maximum width Stqfan et al, 1991, 1992; Janes et al, 1992; Squyres et al, of approximately 2000 km and reaches depths of 15 km below 1992a], radially fractured domes known as novae [Head et al, the mean planetary radius ( km) Multiple ridge and fracture belts are located in the east central portion of the basin 1992], and the sources for large flow fields, such as Mylitta and These structures are up to 1000 km long and elevated no more Kaiwan Fluctus, that are similar in scale to many terrestrial flood basalts [Magee Roberts et al, 1992] The number of than 1 km above the basin floor They are interpreted to be volcanic centers and the range in both scale and character of compressional in origin, the result of horizontal shortening volcanic output from these centers are distinctive from other and crustal thickening [Senske et al, 1991; Solomon et al, 1992; Squyres et al, 1992b] Extensive volcanic plains, rifts on Venus In this paper we examine the role of rifting in generally featureless and radar-dark, have resurfaced the basin the generation of melt through an analysis of the details of the Lada rift structure, the nature and distribution of volcanic and embayed many of the ridge and fracture belts Stratigraphic relationships indicate that local clusters of small shields less centers along the rift, and the relative ages of rifting and than 10 km in diameter are associated with the oldest plains volcanism We consider models for the origin of volcanism units and may represent dis ibuted sources for the earliest and rifting in this region in light of recent studies by plains of Lavinia Otherwise, obvious sources for these plains Bindxchadler et al [1992] on the Origin of lowland plains and are lacking within the basin; local volcanic centers are by Magee Roberts and Head [1993] that linked the formation of concentrated about the periphery of the basin within the rift large-scale flow fields associated with coronae to the presence along the Lada Terra border Lavinia is characterized further by of preexisting rift zones We compare our results to studies of the formation of terrestrial flood basalts and the relative a-10 m geoid anomaly and a -30 mgal gravity anomaly [Bindschadler et al, 1992] Bindschadler et al [1992] interpret Copyright 1995 by the American Geophysical Union Paper number 94JE /95/94JE the basin topography, presence of compressional ridge and fracture belts, lack of numerous large-scale volcanic centers and the presence of gravity anomalies to be the result of an earlystage mantle downwelling or "cold spot" directly beneath Lavinia

2 , : 1528 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT 330 I 0 30 I I '",f :i; " ½ : : : ' -4O i:::!' '? - D One ;" '"":::: ß, ß -:,:,- -20 : ; '-:-- ' ::,,!:'!; ";':e"';;f:;"": :' Ev: e, --, -: : ' :' : Lada T'e"4:':- } Carpo amfana::::: -: ' : :,:,: Se'!u -= ' v '*4;-; :'? ' ½ ::;' ' *: ": ' '},: :, ',:;::::;: : ß :'/ :, '::? :"': :;:,aiwan; :,,';" <j---40 _ - :- -60,,:,, -:-,:, ' e< *<- %;;:"':':: '";: :: '""" ' *':' ;"':½ '::'Uyli, a, -**---'-" * :,, :, ::: : : ; :: ::' s ;"::: -' ;; : : :" % ':: 7i:%:::: '':' : :- '" ' ' :,,: ER"hi "o :, ;:::, :': ':": ;:'"'* '" ; : * " :: '":e:: ' ':" ":*: *- 5'% "; I I ::: :::::,s:: x:: :,:: : ß I 30 Figure 1 Magellan altimetry map (Mercator projection) showing regional topography of the Lada Terra-Lavinia Planitia region Arrows mark location of several volcanic centers along the Lada rift Higher elevations are shown in darker tones; lower elevations in lighter tones Contour interval is 400 m Height accuracy of the Magellan altimeter was better than 50 m; footprint size was <25 km in the high-latitude regions of Venus Region of lower resolution indicates data obtained by the Pioneer Venus Orbiter (where Magellan data are lacking) Modified from Magellan altimetry photoproduct GTDRP3; 2 Data were compiled by the Center for Space Research at the Massachusetts Institute of Technology Lada Terra is an expansive highland region whose boundaries extend from near the south pole north towards Alpha Regio and from approximately 280 ø to 140 ø east longitude [Senske et al, 1991] (Figure 1) Although occasional peaks rise over 2 km above the datum, the region is characterized by typical elevations less than 05 km This is much less than elevations typical of other highlands on Venus, including Beta Regio, Ishtar Terra, and Aphrodite Terra Its boundaries are for the most part gradational with surrounding plains, except along the border of Lavinia Planitia, which is characterized by a series of troughs that comprise the Lada rift Structure and Volcanic Characteristics of the Lada Rift The Lada rift was first mapped as a linear deformation zone and interpreted to be extensional in origin on the basis of its topographic and structural characteristics and its association with volcanic sources [Campbell et al, 1991; Senske et al, 1991; Roberts et al, 1992; Baer et al, 1994] The rift trends almost 3000 km east-northeast along the boundary with southern Lavinia from approximately 625øS, 312øE to 585øS, 7øE; this portion of the rift is named "Kalaipahoa Linea" From here it turns north, following the eastern border of Lavinia approximately 3800 km to (265øS, 359øE) at the southern edge of Alpha Regio (Figures 1, 2); this portion of the rift is tentatively named "Molokai Linea," pending approval by the International Astronomical Union (IAU) The rift is characterized by a discontinuous chain of topographic troughs with outer rises elevated-05-2 km above the surrounding plains A series of 38 topographic profiles compiled from Magellan altimetry were taken across the rift axis; selected profiles are shown in Figures 2 and 3 The rift has a mean width of 160 km, ranging from 35 to 560 km (Figure 4a) The depth ranges from 05 to 42 km, with a mean value of 13 km (Figure 4b) Slopes of the rift walls were measured directly from topographic profiles and found to have an average value of 27 ø (Figure 4c) The greatest slopes are associated with the deepest portions of the rift The east-trending portion of the rift (Kalaipahoa Linea) is composed of multiple subparallel lineaments interpreted to be fractures, scarps, and graben that are approximately 250 to >1000 km long and spaced 2-40 km apart [Campbell et al, 1991; Senske et al, 1991; Roberts et al, 1992] These structures are not confined within the rift troughs nor bounded by distinct scarps They occur in zones of relatively intense to more diffuse fracturing km wide, respectively Regional extension and the emplacement of subsurface dikes are the probable causes for the formation of these linear structures Similar features are observed in the north-trending portion of the rift (Molokai Linea) where the structure of the rift is complicated by a greater abundance of tessera blocks and volcanic centers than in the east-trending portion Along-axis variations in rift topography and structure are plotted in Figure 5 The rift trough is generally narrower and shallower at volcanic centers and wider and deeper in regions of tessera In fact, as reported by Baer et al [1994], in some regions, the rift is virtually indiscernible in the Magellan

3 MAGEE AND HEAD: ROLE OF RIFFING IN THE GENERATION OF MELT 1529 the main rift (Figure 3, profile A-A') An extensive flow field is associated with Eve that extends downslope into Lavinia, branching around and embaying nearby lineament belts These flows appear to have emanated from fractures within the corona annulus Later flows then flooded and emerged from the central portions of the corona Many of the radial fractures observed within the corona annulus are interpreted as dikes that have intersected the surface and fed eruptions of at least portions of the flow field (in a process described by Parfitt and Head [1993b]) Fractures feeding flows are particularly evident in the northwest section of the corona annulus (Figure 6a) The flows are typically km long and 5-25 km wide They are radar-bright in their proximal portions; some of the flows darken distally The earliest flows, as observed in the northwest region of the flow field, are radar-dark and sheetlike in morphology, possibly the result of degraded and smoothed surface textures The entire flow field is approximately 368 x (-28, 3545) (-28, 23) ß" ß i/gi ß **r--,,,:< '-::**: volcanic center liiii!!iiiiiiiiiiii!i!iii : t tessera I ' 2 Z-- lineament belt Figure 2a Mosaic of Magellan radar images C2-MIDRP 30S026;1 and 60S033'1 Projection is sinusoidal Lavinia Planitia"' altimetry However, zones of fracturing are clearly observable all along the rift axis, including regions between volcanic centers along Molokai Linea where there is little topographic expression of the rift structure Variations in rift topography are most likely due to local variations in the amount of thinning and extension along the rift, crustal and lithospheric thicknesses, the distribution of melt, underplating and volcanic flooding and construction Thirteen volcanic centers are located on or near the Lada rift, spaced 90 to over 1400 km apart (Figure 2) The characteristics of volcanic centers along and near the rift are summarized in Table 1 Near the northern-most portion of the rift is Eve, a corona 400 km in diameter (Figure 6) Eve is characterized by a complex annulus of radial and concentric fractures that appear to have formed at least partly within a block of tessera, possibly a fragment of Alpha Regio Within the annulus, a fine-scale radial pattern of closely spaced fractures and graben is cut by larger concentric graben which are superposed by radial graben of similar width oriented north-northwest, aligned with the trend of the rift Some of the graben end in pit chains and may represent subsurface dike emplacement The interior of the corona is heavily flooded by radar-dark lavas A few small shields km in diameter are located within the corona interior A visible trough 120 km long and approximately 7 km wide cuts the corona interior and is located slightly west of 432) Lada Terra Saxpanitum 500 km (-65, 3486) (-65, 23) Figure 2b Sketch map of the north-south trending portion of the rift (Molokai Lines) shown in Figure 2a Black lines represent fractures, scarps, and graben along the rift Heavy black lines indicate locations of profiles given in Figure 3 Dotted line maps boundary of diffuse radar-dark material about the corona Selu Solid black circles represent impact craters; smaller dots represent clusters of small shields (near proximal portion of Kaiwan flow field) Arrows indicate flow direction within flow field; flow field boundaries are outlined in black #

4 1530 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT Figure 2c Mosaic of Magellan radar images C2-MIDRP 60S333;202 and 60S333;201 Projection is sinusoidal 105 km 2 in area An adjacent flow field, known as Eriu Fluctus (Figure 2b), appears to have emanated from fractures extending to the southeast from Eve and may comprise part of the total flow field associated with Eve, increasing the area to 433 x 105 km 2 The geologic history of Eve has been summarized by Squyres et al [1992a] as follows: (1) uplift and doming of the crust and lithosphere and formation of radial fractures, (2) largescale volcanism and the emplacement of an extensive flow field, and (3) central subsidence and concentric fracturing within the annulus Our observations indicate that volcanism and the formation of graben oriented north-northwest have continued past annulus formation, consistent with the observations reported by Baer et al [1994] The fact that interior troughs and the youngest graben are aligned with the trend of the rift suggests that rifting has been active in the latest stages of the evolution of Eve, and may possibly postdate it Carpo is a radially fractured dome or nova [Head et al, 1992] located -660 km south of Eve along the rift (Figure 7) It is 200 km in diameter and dominated by closely spaced fractures and graben radial to the central region Fractures with nonradial trends, some parallel to the trend of the rift, are also observed The radial fractures have been interpreted as the result of lithospheric uplift and doming [Squyres et al, 1992a], although many may be due to vertical and lateral migration of magma from a large reservoir and the emplacement of extensive dikes in the subsurface [Pa 'fitt and Head, 1992, 1993a; Parfitt et al, 1992] The presence of nonradial fractures (open arrow, Figure 7b) indicates that the formation of Carpo may have been "superimposed on a broad pattern of east-west extension" (-53, 30) (-53, 03) km =--' L P!anitia / '/-'- :-/ _- '--Z- Y htta 1 / / c-- /- ' 500 X--_ avinia -::":-:' \ ' \\ \ \ ''i': ', :,:: '-- (-6?5, 310) (-6?5, 05) Figure 2d Sketch map of east-westrending portion of the Lada rift (Kalaipahoa Linea) shown in Figure 2c Dotted lines represent lineaments of uncertain origin Dash-dot lines represent ridges Other symbols are the same as in Figure 2b Blank triangular region represents data gap visible in Figure 2c

5 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT [f 1 [11,,,!,,,I,,,I,,, A B' 05 corona interior has been heavily flooded and contains a cluster of small shields, each <4 km in diameter Closely spaced radial fractures extend from the center of the corona, but most are partially buried by interior flooding and truncated by the corona annulus Some of the radial fractures, aligned with the trend of the rift, cut the northwestern part of the annulus No extensive flow field, save that associated with interior flooding, is associated with Tamfana The geologic history of Tamfana is interpreted to have consisted of radial fracturing followed by subsidence, annulus formation and interior flooding, and continued radial fracturing and/or extension along the rift Approximately 660 km south of Carpo and Tamfana is Selu, a corona 375 km in diameter (Figures 2 and 8) Selu is characterized by domical topography (Figure 3, profile C-C') and a pair of annuli that consist of fractures, scarps, and graben D' L D' I _- " J F' I' ' I [ ' ' I ' ' ' I ' ' '! ' ' ' I ' ' ' distance (km) Figure 3a Sampling of topographic profiles taken across the Lada rift Data were obtained from Magellan altimetry Profile lines shown in Figure 2 Arrows indicate approximate center of narrow rift troughs; double arrows indicate margins of broader troughs (maximum width shown) In profile L-L', letter S indicates the location of the northern border of the Mylitta shield (situated primarily within Lavinia Planitia) and C indicates the shield caldera (located within the rift) I 5 25 d, [ ' c [Squyres et al, 1992a] Thus, extension along the rift may have been active during the formation of Carpo Lobate flows extend approximately 250 km into Lavinia from the periphery of Carpo, forming a relatively small flow field 379 x 104 km 2 in area Immediately to the east of Carpo is Tamfana, a corona 340 km in diameter (Figure 2) The annulus of Tamfana is composed of generally concentric fractures, graben, and ridges The 15 05' ' ' ' I ' ' ' I ' ' ' I ' ' ' I ' "' I ' ' ' I distance (km) Figure 3b (continued)

6 1532 MAGEE AND HEAD: ROLE OF RIFFING IN THE GENERATION OF MELT 45 55t! 1 '1'''1'''1''' relationships, the bulk of volcanism associated with Selu is interpreted to have occurred early in its history An extensive region of diffuse, radar-dark material surrounds Selu and obscures the proximal portions of the flow field (Figure 2) The origin of this material is unclear, but may be a soil layer produced by impact processes or, less likely, during pyroclastic N' 05-1 (a) 6- mean = 159 km 2- _ edian = 148 km width (km) (b) ½ m,;, i::::::: '!,,,,,' ß mean= 13 km median = 10 km El0 r," [ =: '[, ½::::::::: _ ' 0 ' ' ;0'0;,00 ' I ' I ' I ' ' 1 ' distance (km) Figure 3c (continued) l'"?g1 t:: :"='4' ;':! -*:===,:,:,::½'':,::: g i i i i i i depth (kin) These structures define concentric annuli but are generally oriented northwest, parallel to one another and aligned with the trend of the rift As a whole, the annulus of Selu is not as well developed as that of Eve and other coronae along the rift, suggesting Selu may be at an earlier stage in its evolution, based on published models of corona evolution [Pronin and Sto/hn, 1990; Stofan and Head, 1990; Stqfan et al, 1991, 1992; Janes et al, 1992; Squyres et al, 1992a] The interior of the corona is dominated by radial fractures, graben, and pit chains, some of which extend several hundred kilometers up to 1480 km beyond the corona annulus The longest of these structures do not maintain a strictly radial orientation but curve to align with the trend of the rift, consistent with the suggestion that Selu formed in a preexisting east-west tensional stress regime [Squyres et al, 1992a] A flow field 608 x 105 km 2 in area extends from Selu west into Lavinia and, to a lesser extent, east into local lows within Lada Terra (Figure 2) The lavas appear to haverupted from the radial fractures Figure 4 Histograms and graben, the majority of which may now be buried beyond average the periphery of the annulus Based on crosscutting topographic mean = 27 ø rnedian =17 ø slope (degrees) of range in (a) average rift width, (b) depth, and (c) maximum slope Based on 35 profiles obtained from Magellan altimetry data

7 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT 1533 a) Kalaipahoa Molokai Linea ß 6OO b) 6054 N S/E W c) N S/E W Figure 5 Along-axis variations in rift dimensions Values were obtained from a series of 38 altimetric profiles (each tick on the horizontal axis represents one profile) taken perpendicular to the strike of the rift In each plot, the traverse is down the rift along Molokai Linea from near Eve (left) to where the rift changes orientation near Eithinoha to near its westernmost portion along Kalaipahoa Linea (right) Gaps in the Magellan altimetry data preclude measurements of the westernmost portion of the rift Vertical bars mark locations of volcanic centers and tessera blocks; Ev, Eve; C/T, Carpo/Tamfana; S, Selu; T, tessera block; K, Kaiwan; Ei, Eithinoha, N, unnamed nova; M, Mylitta; J, Jord Horizontal lines represent mean values (a) Rift width; solid line represents width of deformation zone, dashed line represents width of topographic trough (b)rift elevation; solid line represents elevation of rift crest, dashed line represents elevation of rift floor (c) Rift depth (relief between crest and floor)

8 1534 MAGEE AND HEAD: ROLE OF RIFLING IN THE GENERATION OF MELT Table 1 Volcanic Centers along and near the Lada Rift Zone Volcanic Location, Diameter, Type or Structure Flow Field Area, Center (latitude, longitude) km of Edifice km 2 Eve (-32, 359) 400 concentric corona a x 105 Carpo (-375, 3) 200 nova b 379 x 104 Tamfana (-363, 6) 340 concentric corona a Selu (-425, 6) 375 radial/concentricorona 608 x 10 5 Kaiwan -(-488, 98) large flow field with 260 x 105 Fluctus no obvious central source Derceto (-468, 202) 200 concentric (asymmetric?) 508 x 105 corona/caldera a Sarpanitum (-523, 146) 170 concentric corona/caldera a Eithinoha (-573, 82) 510 concentric corona a 113 x 105 unnamed (-585, 0) 115 nova b 337 x 103 nova Mylitta (-583,3514) central shield large flow field with 30 x 105 F!uctus -200 asymmetric central shield Jord (-585,3495) 190 radial/concentric corona a? Kamui- (-633,322) 300 concentric/double ring Huci corona Quetzalpetlatl (-67, 0) 800 asymmetric corona a 147 x 106 with long, radiating fractures Only the names Eve, Kaiwan Fluctus, Mylitta Fluctus, Eithinoha, Kamui-Huci, and Quetzalpetlatl have been accepted by the IAU; all others are provisional aclassification of corona types from StoJ tn et al [1992] /'radially fractured dome; defined by Head et al [ 1992] eruptions, and windblown from more distant regions It is very al, 1987] and many radial fracture systems on Venus similar to material observed near the proximal portions of interpreted to be dike swarms have mapped lengths of similar Mylitta Fluctus [Magee Roberts et al, 1992] magnitude [Parfin and Head, 1992] It is also possible that Kaiwan Fluctus is a flow field 26 x 105 km 2 in area located dikes and fissures emplaced within or parallel to the rift served approximately 875 km to the south of Selu along the trend of as conduits for the Kaiwan lava flows Local flooding and the the rift (Figure 9) Kaiwan is very similar in scale and presence of several clusters of small shields within the rift and morphology to the flow field of Mylitta Fluctus, located on the near the proximal portion of the flow field suggesthat fissure east-west trending portion of the rift [Magee Roberts et al, vents may have been located in this area However, this 1992] Unlike Mylitta and other large flow fields along the portion of the rift is not breached or superposed by the flow rift, however, Kaiwan lacks an obvious central source region field and is associated with an outer rise that exhibits Lava flows within Kaiwan may have originated from fissures m of relief (Figure 3, profile F-F') Perhaps fissure vents that were obscured by subsequent burial The flow field is cut by associated with the emplacement of Kaiwan were located to the several of the longest fractures or dikes associated with Selu; west of the main rift and subsequently buried The relative other fractures are buried by the flow field It is possible that timing of the emplacement of Kaiwan and extension along the flows within Kaiwan may have erupted from dikes emplaced rift is unclear from Selu This would require a very large volume of melt to Sarpanitum is 170-kin-diameter circular structure located have migrated -875 km laterally before rising vertically to within a block of tessera along the rift approximately 400 km erupt at the surface Although perhaps unusual, the possibility southeast of Kaiwan (see Figures 2 and 9) It is defined by of such an event is supported by the fact that lateral dike concentric fractures and has been interpreted as a possible emplacement over 1500 km from a central source has been caldera or concentric corona by Stofan et al [1992] The documented for the MacKenzie dike swarm on Earth [Gibson et interior is approximately 15 km deep (Figure 3, profile G-G')

9 MAGEE AND HEAD: ROLE OF RIb-TING IN THE GENERATION OF MELT 1535 Figure 6a Magellan image of Eve and is flooded by radar-dark lavas No exterior flow field is obvious Graben associated with deformation within the rift or tessera block cut portions of the annulus No age relationship between Sarpanitum and the rift is clear Eithinoha is a corona 510 km in diameter located approximately 570 km from Sarpanitum at the point where the rift bends from a generally north-south to an east-west orientation (Figure 2) It has a relatively narrow annulus of concentric and radial ridges, scarps, and graben (Figure 10) A 75-km-wide trough oriented N40øE and approximately 15 km deep bisects the corona and cuts the northern portion of the annulus (Figure 3, profile I-I'; Figure 10) The trough is located within a broader zone of fractures km wide The presence of the trough indicates that Eithinoha may have extended due to continued activity along the rift A second set of fractures 150 km wide is oriented perpendicular to the trough and truncated by the corona annulus Lava flows have emanated from the trough to flood the interior of the corona; some are cut by the annulus Additional flows erupted from concentric and radial fractures within the annulus and form a flow field approxi nately 113 x 105 km 2 in total area Portions of the flow field may have been buried by volcanic plains within Lavinia Fractures and graben trending east-west along the rift splay to the south about the periphery of Eithinoha (Figure 10), indicating that activity along this portion of the rift was modified by the presence of Eithinoha or stresses linked to its formation Eithinoha may have formed prior to or contemporaneous with the development of this portion of the rift The geologic history of Eithinoha appears to have been similar to that of Eve and Tamfana, with a greater degree of deformation related to subsequent extension along the rift Between Selu and Eithinoha the rift cuts through two large blocks of tessera approximately 400 by 900 km in width In (-286, 3530) (-286, 10) Lavinia P!anitia (-357, 3533) (-357, 2,0) Figure 6b Sketch of Eve from image in Figure 6a Solid gray pattern represents flow field, which is comprised of radar-bright digitate flow lobes and radar-dark sheetlike flows Flows appear to have erupted from radial fractures (interpreted as dikes) observed within the corona annulus, although flow emplacement probably occurred both before and after annulus formation Later flows have erupted from the interior of Eve Black arrows indicate flow direction Digitate flows do not extend north of the dash-dot line; the flow field is radar-dark in this region White arrows mark radial north-trending fractures that end in pit chains and are superposed on concentric fractures (see text for discussion) Dotted line maps the boundary of diffuse radar-dark material interpreted as the scar of an impact event produced by a meteoroid that was destroyed before it reached the surface, as discussed by Phillips et al [1991] Data gaps are shown by vertical bars Other patterns and symbols are the same as in Figure 2

10 o 1536 MAGEE AND HEAD: ROLE OF R1FHNG IN THE GENERATION OF MELT Figure 7a Magellan image of Carpo (-380, 225) (b) (-380, 140) ; '"' '" '*" - ; /: i:q, this region, troughs along the rift exhibit km of relief and vary in maximum width from 170 to 270 km (Figure 3, profiles D-D'- H-H') These troughs are typically deeper and narrower than elsewhere along the rift (Figure 5) Approximately 420 km to the west of Eithinoha is an unnamed nova 115 km in diameter (see Figures 2 and 10) This feature is dominated by closely spaced fractures which radiate from an interior circular structure 50 km in diameter A few :d **' :;;el' ::,;4,;,,''1 lobate flows that appear to be associated with this edifice extend into Lavinia, forming a flow field 337 x 103 km 2 in (-3585, 3597) (-3585, 38) s?? :; (-455, lo) (-455, 140) Figure 8 (a) Magellan image and (b) sketch of Selu The proximal portion of the extensive flow field to the southwest of Selu is obscured by diffuse radar-dark material and cannot be extended with certainty to the corona annulus, although the corona is interpreted as the source of the flow field (More distal portions of this flow field are shown in Figures 2 and 9) Open arrows mark a fracture that changes in orientation from being radial to the center Of Selu to aligning parallel to the Lada rift As discussed in the text, this is evidence that Selu may have been emplaced in a zone of regional extensional Patterns and symbols same as in Figure 6b area The rift cuts through just to the north of the edifice, exhibiting about 15 km of relief (Figure 3, profile J-J') Mylitta Fluctus is perhaps the most distinctive volcanic center along the rift, located about 490 km to the west of the (-3 75, 3597) (-3875, 39) nova mentioned above Mylitta is characterized by an ig,r, Ii, Sketch of Carpo from image in Figure 7 Digitate asymmetric shield 200 km in radius with an elongate caldera flow field appears to ha e erupted f om fractures (dikes) radial to km in diameter located within the rift and a superposed th center of Carp; uncertainties th boundary of the flow extensive flow field that extends up to 1000 km north into Lavinia and covers an area of approximately 3 x 105 km 2 flooding is observed at Carpo Open arrow marks the location (Figures 2 and 11) Portions of the flow field were emplaced of, p orni m o ai l fr ctur o,iet a parallel to the rift south of the shield The detailed morphology and stratigraphy As discussed in the text, the presence of nonradial fractures of Mylitta were described previously by Magee Roberts et al indicates Carpo may have formed in the presence of a tensional [1992] Flows associated with Mylitta are superposed on stress regime related to the Lada rift Patterns and symbols same fractures along the rift and show little evidence for subsequent as in Figure 6b deformation by continued rifting (except possibly by the

11 ß MAGEE AND HEAD: ROLE OF RIFIING IN THE GENERATION OF MELT 1537 :::-; ;:;, ::* ;:-",',, :-- :-r,';:- ;*--:; ;?-' :;-- ',' ;';*; :;: ;; :::,,-: -- :,,%-,',½-,,:** :: Figure 9a Magellan image of Kaiwan Fluctus presence of arcuate scarps within the shield that are subparallel to the trend of the rift [Magee Roberts et al 1992]) Given that the majority of volcanic edifices along the riff show evidence for rift-related deformation and Mylitta does not, Mylitta may be one of the youngest volcanic centers along the rift, or at least one of the youngest centers relative to rifting in the Lada Terra region hmnediately to the west of Mylitta is the corona Jord, approximately 190 km in diameter (Figure 11) Multiple concentric fractures and graben define the corona annulus, which is both cut by fractures and graben within the rift (some radial to the corona interior) and embayed by flows associated with Mylitta [Magee Roberts et al, 1992] Numerous lava channels, pits, and small shields 2-12 km in diameter are located near the center of the corona No flow field clearly associated with Jord can be distinguished from that of Mylitta The last major volcanic center located along the rift is Kamui-Huci, a double ring corona 300 km in diameter approximately 1420 km west of lord(figures 2 and 12) The double annulus is comprised of concentric fractures, scarps, and (-450, 3593) (-450, 148) ß ß : I Lavinia P!anitia" / I km (-526, 3570) (-526, 150) Figure 9b Sketch of Kaiwan Fluctus (and portions of Selu flow field) from image in Figure 9a Kaiwan flow field may extend to dashed line Open arrows mark locations of fractures that appear to represent subsurface dikes that have been emplaced laterally from Selu, some of which may have fed portions of the Kaiwan flow field Other possible sources for the Kaiwan flow field are fractures and small shields (dots) distributed within the rift Patterns and symbols same as in Figure 6b

12 1538 MAGEE AND HEAD: ROLE OF RIFFING IN THE GENERATION OF MELT (a) defined by fractures that bifurcate and extend approximately 400 km from the western portion of the corona annulus (Figures 2c, 2d) Gaps in both image and altimetry data prevent an exact determination of the western extent of the rift (Figures 1, 2) It is possible that the rift may continue further to the westnorthwest, encompassing a greater portion of the perimeter of Lavinia Planitia At the time of this writing, no topographic data were available for the portion of the rift along Kalaipahoa Linea near ( ) (-540, 00) (-540, 180) (a) (-6175, 35675) (-6175, 190) Figure 10 (a) Magellan image and (b) sketch of Eithinoha and the unnamed nova Volcanism associated with Eithinoha has produced flow fields to the northwest and east of the corona annulus and has flooded portions of the corona interior These flows appear to have erupted from fractures (dikes) within the corona interior and annulus Peripheral shields km in diameter with central pits and/or calderas are indicated by open circles Open arrows mark the location of fractures that splay about the southern boundary of Eithinoha and those within a central trough (in line with the trend of the 'Lada rift) that cuts the corona Patterns and symbols same as in Figure 6b (-5226, 3455) (-5226, 3574) "":' J i-: i: '-:;!':?-: 5:-:'::'-'i::: '-,'?'"'- '"" -'""" '" " ' q':'?>: :': ' "'! "' ' '--"" '"' :" ' Shi, graben These structures are cut by additional fractures and graben, some of which are oriented radial to the corona interior and others parallel to fractures associated with the rift No distinctive flow field is associated with this corona although mottling in the surrounding plains suggests that a highly,, : a?: / degraded flow field may exist By assuming the corona was (b) originally completely circular, Senske et al [1991] estimated it may have been extended by about 18% of its diameter by continued deformation along the rift Their analysis was based on radar image data obtained by the Arecibo Observatory in Figure 11 (a) Magellan image and (b) sketch of Mylitta Fluctus and Jord Dotted lines mark boundary of radar-dark diffuse material near proximal portions of Mylitta Fluctus Most of the Mylitta flow field is superposed on the shield and 1988 Recent Magellan data do not reveal any obvious has flowed north into Lavinia; a portion was emplaced to the deviation from circularity, although deformation due to rifting is indicated by crosscutting relationships of fractures within south Jord has been extensively embayed by Mylitta See text and Magee Roberts et al [1992] for details Patterns and the rift and corona annulus The western limit of the rift is symbols same as in Figure 6b

13 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT 1539 Derceto is an elongate corona or caldera structure km in diameter located approximately 670 km east of Kaiwan Fluctus (Figure 2) An extensive fan-shaped flow field [Lancaster eta/, 1992] has been emplaced east of the Lada rift Additional material flowed north and west into local lows within tessera along the rift The total area of the flow field is 508 x 105 km 2 No clear age relationship between rifting and this center is observed To summarize, 13 volcanic centers of varying morphologies and an average center-to-center spacing of about 630 km are located along or near the Lada rift Little or no volcanism is evident elsewhere on the rift Crosscutting relationships indicate that rifting has been active prior to and since the formation of most of the volcanic centers Mylitta appears to be the youngest center along the rift, as there is little evidence for substantial deformation subsequent to its formation No systematic progression in age or scale of eruptive products is (-6085, 31575) (b) (-6085, 32675) observed among the volcanic centers along the rift Three broad categories of rifts have been recognized on Venus [Senske, 1992; Senske and Head, 1992] The characteristics of rifts from each of these categories are given in Table 2 As described by Senske [1992] and Senske and Head [1992], simple rifts, such as Devana Chasma in Beta Regio, are characterized by single, continuous linear troughs that reach depths of 1-5 km They are thousands of kilometers long, km wide, and bounded by steep scarps which define the extent of fracturing The rift flanks show little deformation and may be elevated as much as 3 km above the surrounding plains Simple rifts occur near or within volcanic rises and often radiate from a large central edifice such as Theia Mons in Beta Regio (-660, 31575) and Ozza Mons in Atla Regio Volcanism is concentrated Figure 12 (a) Magellan image and (b) sketch of Kamui-Huci primarily at the central edifice Extension associated with corona near the western end of the Lada rift Dotted lines map simple rifts is thought to be the result of regional uplift due to lineaments of uncertain origin No flow field is mapped, the upwelling of a mantle plume [Senske, 1992; Senske and although mottling within the plains to the south of the corona Head, 1992; Senske et al, 1992] (refer to image) suggests a flow field associated with this center Corona chains are characterized by multiple coronae may have been emplaced and subsequently degraded Patterns connected by zones of intense fracturing 500 to over 1000 km and symbols same as in Figure 6b wide [Senske, 1992; $enske and Head, 1992] Parga and Hecate Chasmata contain over 75 coronae [Stofan eta/, 1993] with an Kamui-Huci Image and altimetry data available west of Mylitta (Figures 2 and 3, profiles M-M' - Q-Q') indicate riff-related troughs vary in width from 40 to 225 km and exhibit km of relief The rift cuts through several irregular blocks of tessera and is associated with a broader, more diffuse zone of fracturing up to 630 km wide Two additional volcanic centers are located near but not directly along the trend of the riff Both are associated with substantial amounts of volcanism Quetzalpetlatl is one of the largest coronae on Venus and is approximately 800 km in diameter [Solomon et al, 1991] It is located south of Mylitta at a center-to-center distance of 1200 km Linear fractures and graben interpreted to represent subsurface dike emplacement [Magee Roberts et al, 1992] extend up to 530 km between Mylitta/Jord to the annulus of Quetzalpetlatl, where they are buried by lava flows Quetzalpetlatl has one of the largest measured flow fields on Venus, covering approximately 147 x 106 km 2 Tectonic and volcanic activity associated with Quetzalpetlatl are interpreted to at least partially postdate activity along the Lada rift [Magee Roberts et al, 1992] Comparison With Other Venus Rifts average center-to-center spacing of about 430 krn Many of the coronae are contiguous to one another Corona chains are typically many thousands of kilometers in length and found almost exclusively in plains regions, often extending from a major volcanic rise [$enske, 1992; $enske and Head, 1992; Sto fan et al, 1993] They are generally surrounded by arcuate troughs and steep scarps with slopes ranging up to 10 ø to greater than 30 ø [Ford and Pettengill, 1992; Senske, 1992; Senske and Head, 1992] Zones of most intense fracturing are restricted to the topographic lows along the rift More diffuse zones of fracturing occur within 300 km of the main trough [Hamilton and Stofan, 1993; Stofan et al, 1993] In contrast to simple rifts like Devana Chasma, corona chains are complex branching networks with several discontinuous offsets along strike [Stofan et al, 1993] Corona chains are interpreted to represent zones of extension along which mantle diapirs have risen to form coronae [Stofan et al, 1993] The origin of these zones of extension remains unclear Fracture belts, such as the Lada rift and those observed along the southern margin of Aphrodite Terra and within Aino and Sedna Planitiae, are characterized by a discontinuous chain of

14 1540 MAGEE AND HEAD: ROLE OF RI TING IN THE GENERATION OF MELT Table 2 Characteristics of Venus Rifts Rift Location Devana Chasma Beta Regio Width, Depth, Length, a a a Volcanic Associations Structural Characteristics Theia Mons; large central edifice "simple rift; "a intense fracturat juncture of three rift arms a ing restricted to topographically distinct trough; branches edifice volume approximately into three rift arms; flanks 144, ,000 km 3 a elevated km above surrounding plains; little no other significant volcanics a deformation of rift flanks a 20-30% extension a,/' [ Guor Linea Western Eistla Regio 45_75t', c 07_19 c 1000 t' little volcanism except at Gula Mons c "simple rift; "a topographic trough defined by two major scarps km in relief flanks elevated km above regional topography c Ganis Chasma Aria Regio up to 300 t' t' 1000 t' volcanism concentrated at Ozza and Maat Mons, major edifices at juncture of multiple rift arms c "simple rift; "a one of five major rifts radial to Ozza Mons Parga and 500 to 8000 to Hecate 1000 a > 10,000 d Chasmata Lada Terra >6000 (160 avg) (13 avg) asenske [1992] t'solomon et al [1992] CSensk et al [1992] dstofan et al [1993] eharnilton and Stofan [1993] fcalculated from Campbell et al [ 1991 ] gmagee Roberts et al [1992] multiple (>75 /) coronae with an average center-to-center spacing of-430 km; avg diameter-265 km; avg flow field area -83 x 104 km 2 multiple (11-13) volcanic centers including coronae, edifices, and large flow fields; spaced -630 km apart on average; avg diameter km; avg flow field area- 19 x 105 km 2 "corona chains;"a composed of coronae connected by intense, complex zones of fractures that often branch and conain several discontinuous offsets along strike; more diffuse zones of fracturing occur within 300 km of the main trough a'd'e 10-15% extensionf [ "fracture belt;"a discontinuous array of fractures and troughs along a highland/lowland boundary; fractures not confined to any distinct trough bounded by well-defined scarps; contained within broad topographic rise km above surrounding plains 12 (average value for rift) topographic lows and relatively broad zones of fracturing fracturing are not confined within a trough bound by steep [Senske, 1992; Senske and Head, 1992] As described earlier, scarps It has an average depth of -1 km, much shallower than the Lada rift is located at the margin of a large highland, along simple rifts like Devana Chasma, where 3 to 5 km depths are the border of a relatively circular basin It is not one of several common, and it has an average slope (27ø), much lower than rift arms radiating from a major volcanic edifice atop a regional that typical for either simple rifts or corona chains Like rise, nor does it branch in any complex fashion Zones of corona chains, the Lada rift is associated with multiple volcanic

15 MAGEE AND HEAD: ROLE OF RIFFING IN THE GENERATION OF MELT 1541 (A) PRIOR TO ONSET OF DOWNWELLING: mantle litho- sphere Lavinia Planitia crust ' _ Lada Terra [:i!i:i,,71 ' '-" AFTER DOWNWELLING INITIATED: (B) peripheral j extension =rdec nc ta l d 9 thickening due to coupling of lithosphere to mantle flow and subsequent flexure mantle downwelli ng (c) peripheral extension less and increased formation thickening developed/overpfinted to west (v of Lada associated rift and volcanics " ridge belts /-, mantle downwelling upwelling related to mantle return flow and/or passive/buoyant upwelling related to rifting due to peripheral stresses of downwelling Figure 13 Sketch illustrating possible downwelling-related origin of Lada rift and associated volcanism, based on the model of downwelling beneath Lavinia Planitia described by Bindschadler et al [1992] See text for explanation Cross-sections (westo left) not drawn to scale; sketch in Figure 13c is enlarged only to show detail and does not represent change in regional dimensions centers However, these are much fewer in number and rarely southern margin of Aphrodite Terra and the Lada rift may be the contiguous, separated by a larger average spacing of 630 km result of slope failure or gravity sliding Such an origin seems Models for the origin of the Lada rift are considered in the more likely for the Aphrodite fracture belt, where the overall following section difference in topography between the highland margin and adjacent plain locally exceeds 6 km and the slopes are steep, Origin of Rifting than for the Lada rift, where the maximum relief between the highland and Lavinia is about 3 km and the slopes are relatively On the basis of their associations with highland margins, gentle Senske [1992] and Senske and Head [1992] suggested that the An alternative model is linked to the possibility that mantle origin of fracture belts such as the one observed along the downwelling is occurring beneath Lavinia Planitia

16 1542 MAGEE AND HEAD: ROLE OF RIb-TING IN THE GENERATION OF MELT [Bindschadler et al, 1992] A strong coupling of mantle flow to the lithosphere is expected to cause tectonic deformation on Venus due to the inferred absence of a low-viscosity zone or estimate includes a possiblextension of fractures along the rift to 20øS, 350øE, about 1200 km to the northwest of Eve If mantle downwelling has produced extensional strain about the asthenosphere and the presence of high surface temperatures periphery of Lavinia, it is curious that only the margin along that imply shallow levels of ductile deformation within the crust [Bindschadler and Parmentier, 1990; Phillips, 1990] Lada Terra has deformed Assuming the highland crust of Lada is thicker than its surroundings, it might represent a region of Thus, in its early stages, mantle downwelling, as modeled for reduced lithospheric strength As discussed by Grimm and Venus conditions by Bindschadler and Parmentier [1990], Phillips [1990], stresses may be focused at regions of strength results in subsidence, compressional hoop strains and peripheral extensional radial strains These conditions are consistent with the basin topography of Lavinia (its generally circular shape might be due to downwelling of a cylindrical discontinuities Therefore, if the stresses induced about the margin of Lavinia are relatively minor, it is possible that strains would be manifest only in the comparatively weaker crust of Lada Terra, or along the highland/planitia boundary, planform), the presence of interior compressional ridge and the location of a discontinuity in crustal strength fracture belts, and the negative gravity anomaly associated with Furthermore, the west-northwest margin of Lavinia is defined Lavinia [Bindschadler et al, 1992] We propose that the by Themis Regio (the eastern extremity of Parga Chasma, a peripheral extensional strains may also account for the formation of the rift along the adjacent margin of Lada Terra corona chain discussed previously) and the regional volcanic rise of Dione Regio (the site of shield volcanoes Ushas, Innini, (Figure 13) and Hathor Montes) (Figure 1) Both of these areas are To evaluate this model, two questions must be addressed: (1) are the strains sufficient to account for the scale of deformation observed?, and (2) why is the formation of a peripheral rift interpreted as possible regions of mantle upwelling, thermal uplift, and limited extension [Senske et al, 1991; Keddie, 1993; Stofan et al, 1993] The strain associated with mantle about Lavinia restricted to the margin of Lada Ten'a? Estimates downwelling under Lavinia may have been modified or of the time and length scales of crustal deformation related to overprinted by local tectonic activity in these areas In mantle flow tectonics are model dependent and vary strongly addition, the deepest region of Lavinia and the center of the with the depth of associated buoyancy forces, crustal thickness, gravity anomaly is in the eastern portion of the basin, at about viscosity structure, and the presence of elastic layers within the 45øS, 350øE [Bindschadler et al, 1992] The greatest strains lithosphere [Bindschadler and Parmentier, 1990] Nonetheless, associated with downwelling centered below the eastern section strains are expected to be "relatively small" in the early stages of Lavinia may have been produced along its closest margins, of downwelling and are likely to be influenced by regional ie, along the border of Lada Terra stresses and inhomogeneities in lithospheric strength [Bindschadler et al, 1992] If peripheral extensional strains Origin of Volcanism and Relation to Rifting produced solely by mantle downwelling are unable to account for the scale of rifting observed, it is possible that subsequent Unlike simple rift systems on Venus, the characteristics of the Lada rift indicate it has not formed due to tectonic forces dike emplacement oriented by the presence of a tensional stress regime may enhance the strains sufficiently to produce the observed amount of extension A similar situation has been described at Kilauea Volcano in Hawaii Dikes emplaced along Kilauea's east and southwest rift zones often result in extension of the south flank of the volcano Such faulting allows the upwelling related to passive rifting, active upwelling of anomalously hot mantle unrelated to rifting, and mantle flow volcano to adjust to dike emplacement and generates stress related to downwelling are considered below patterns needed to constrain subsequent dike intrusion along the rift axis [Swanson et al, 1976; Dvorak et al, 1986; Dieterich, 1988] As described earlier, much of the deformation Upwelling Related to Rifting along the Lada rift may be due to the emplacement of subsurface dikes Fractures and graben related to possible dikes whose orientations have been modified by the presence of a The fact that volcanism is distributed along the trend of the rift, at a number of centers spaced an average of 630 km apart, suggests that upwelling of the mantle in response to preexisting tensional stress regime are best observed at the lithospheric extension may be a source for melt Melt corona Selu, where fractures initially radial to the corona at close distances (within the annulus) become parallel to the rift generation by lithospheric extension and associated upwelling is thought to be dependent on the degree of extension and the further away, extending for distances up to 1480 km The difference between these two scenarios is that motion along the south flank of Kilauea is driven by dike emplacement For the case of the Lada rift, ongoing regional extension related to downwelling would have oriented dikes emplaced along the rift such that continued dike emplacement contributed to the total amount of extension Thus, if initial strains produced by downwelling forces in Lavinia were insufficient to account for the observed amount of rifting within Lada, dike intrusions oriented parallel to and emplaced within the rift may have produced additional extensional strains that account for the deformation currently observed along the rift Based on our observations and mapping, the Lada rift represents 44-53% of the margin of Lavinia The higher associated with the upwelling of a large mantle plume If it has formed as a result of regional extension, possibly related to mantle downwelling under Lavinia Planitia, what is the origin of the volcanism associated with the rift? The contributions of temperature of the upwelling mantle [eg, White et al, 1987; McKenzie and Bickle, 1988; White and McKenzie, 1989] The amount of extension is often estimated by the areal extension of some portion of the lithosphere and expressed in terms of the ratio of the final to original surface area (or width) of the region being rifted [McKenzie and Bickle, 1988] For example, recent studies have reported that the melting of normal, thermally unperturbed asthenosphere begins when the lithosphere has been extended by [I-25 Melting commences at lesser amounts of extension ([I-12) when the upwelling mantle is øC hotter than normal [Latin and White, 1990; Mohr, 1992] For terrestrial rifts, the amount of lithospheric stretching, or [, is often determined by the use of seismic profiles and

17 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT 1543, \0 0/ \ \ / \ / X2, x = 3a = 3b-3Ax x2 = x + Ax = 3a + Ax = 3b- b = a + Ax Ax = a = b-ax 2Ax 2h tan 0 x2 3b- 2Ax x Bb- BAx b a b b - Ax Figure 14 Sketch showing method for estimating amount of extension along the rift Values for b (crest-to-crest width of rift trough), h (depth from rift cresto floor), and 0 (slope of rift wall) are measured from topographic profiles and used to calculate Ax (change in width or difference between original width a and observed width b) and [ (ratio of final to original width) estimates of thermal subsidence On Venus, no such data are available For the purposes of this study, crudestimates of [3 identical to that estimated by restoring Somerville to original circularity (127), assuming an original diameter of 37 km and were calculated from topographic profiles by assuming a Ax -10 km [Solomon et al, 1992] The total amount of simplistic rift geometry consisting of symmetric low-angle extension along Devana Chasma could be greater than these normal slip faults, as shown in Figure 14 Values for b (crest- estimates if Somerville was emplaced relatively recently in the to-crest width of rift), h (rift depth), and 0 (slope of rift walls) were measured directly from the topography and used to history of the rift; these estimates do not necessarily imply similar amounts of extension for both Devana Chasma and the calculate Ax (change in rift width) and [3 Multiple topographic Lada rift While these results are encouraging, the method profiles over and adjacent to each volcanicenter were analyzed outlined in Figure 14 remains a very approximate estimate of to lessen the chance of underestimating the local rift width due to volcanic infill The average value of [ (using the second formulation given in Figure 14) for the Lada rift is 12; values for various volcanic centers along the rift range from 19 to 12 the amount of extension along any Venus rift Of course, if the amount of extension along the Lada rift was enhanced by oriented dike emplacement, as discussed earlier, then some error will be introduced when using these values to evaluate the (Tables 2, 3) As a test of the relative accuracy of this method, efficacy of passive upwelling in producing the amount of the same measurements were applied to profiles across Devana volcanism observed Chasm and the impact crater Somerville, which has been split Given estimates for the amount of extension along the Lada and separated by % extension along Devana [Senske, rift, can passive upwelling account for the volume of melt 1992; Senske et al, 1992; Solomon et al, 1992] Values observed, for example, at Mylitta Fluctus? The volume of obtained for [ using the two formulations given in Figure 14 extruded melt at Mylitta is -2 x 104 klrl 3 [Magee Roberts et al, range from 11 to 15 The average value, 13, is nearly 1992] Assuming a range of extrusion to intrusion ratios of 1:5

18 1544 MAGEE AND HEAD: ROLE OF RIFIING IN THE GENERATION OF MELT Table 3 Rift Zone Parameters at Specific Volcanic Centers Volcanic Center Width a, km Depth a, km Slope, deg 1 [ 2 Eve Selu Kaiwan Fluctus Eithinoha Mylitta Fluctus Unnamed nova Kamui-Huci 12 t' Parameters 1 and are ratios of final to original width; see text for explanation aaverage width and depth of rift compiled from multiple topographic profiles across volcanic center and across rift immediately adjacento each center; profiles in addition to those shown in Figure 3 were examined btopography data unavailable at time of writing; calculated from 18% extension estimated by Senske et al [1991] to 1:10 [Crisp, 1984], the total volume of intruded melt is 1 x 105 to 2 x 105 km 3 This converts to a thickness of melt beneath the rift of kin, assuming melt has intruded under a portion of the rift 370 km long and 125 km wide (based on the length of rift superposed by flows at Mylitta and the average width of the rift from topographic profiles across this region of the rift) To account for these values of melt thickness intruded below the rift, and the estimate of [3-12, a mantle potential temperature of 1480øC is required (determined various sites along the rift For example, the second largest value of is measured at the unnamed nova, which has the smallest measured flow field along the rift In addition, the value calculated for Mylitta is identical to (not larger than) that of the average rift (12) Finally, the values for Selu, Eithinoha, and Kaiwan are all identical, despite large differences in flow field area This is not consistent with the idea that melt generation by passive upwelling is dependent on the amount of extension, although estimates of at the various volcanic centers may be inaccurate due to flooding and partial burial of portions of the underlying and adjacent rift In addition, it is unclear why passive upwelling should produce so much melt at the Lada rift and not at other, more developed sites of rifting on Venus For example, Devana Chasma is es- sentially barren of volcanism along most of its length; from Figure 22a of McKenzie and Bickle [1988]) Under volcanic activity is concentrated at a central edifice, Theia terrestrial conditions, such a mantle temperature would indicate Mons [Senske, 1992] In fact, at most rifts where significant the presence of a mantle plume or thermal anomaly øC amounts of volcanism are present on Venus there is strong hotter than normal asthenosphere and it would seem that simple evidence for the presence of mantle thermal anomalies passive upwelling of normal mantle is not sufficient to account [Bin&cradler et al, 1992; Grimm aml Phillips, 1992; Senske for the observed volumes of melt at Mylitta However, based et al, 1992; Solomon et al, 1992] on parameterized convection models of Venus, mantle temperatures are thought to be at least --200øC hotter than in These observations do not necessarily rule out rift-related upwelling as a significant componento melt generation along Earth [Stevenson et al, 1983; Phillips and Malin, 1983] Such the Lada rift Factors other than the total amount of extension elevated upper mantle temperatures were predicted to produce a crust 2-3 times thicker in a hypothetical Venus spreading center environmenthan in a typical terrestrial mid-ocean ridge [Sotin et al, 1989] Therefore, it is plausible that passive upwelling of normal Venus mantle might be sufficient to produce the observed volumes of melt, given our estimates of the amount of extension along the Lada rift also have an influence on the volume of melt generated in a rifting environment (Table 4) Higher mantle and lithospheric temperatures may locally enhance melting independently of the total amount of extension [White et al, 1987; McKenzie and Bickle, 1988; White and McKenzie, 1989; Griffiths and Campbell, 1991; Mohr, 1992; Hill, 1991] Decompression melting may be reduced by locally greater lithospheric Is passive upwelling consistent with other observations thicknesses [eg, Hill, 1991] Extension rate is also regarding volcanism along the Lada rift? Consider, for significant in melt production Due to increased rates of heat example, the range of values associated with volcanic centers conduction, slow spreading rates may greatly reduce the amount along the rift and compare these values to the range in flow of melt generated by passive rifting [eg, White, 1992], alfield area There is no correlation between the estimated though the dependence of melt production on spreading rate is amount of extension and the scale of volcanics observed at mitigated in regions of lower mantle viscosities due to the importance of compositionally buoyant mantle upwelling [eg, Sotin and Parmentier, 1989] Extension rates also affect the advection of a thick depleted mantle layer away from the zone of melting such that, at low rifting rates, melting is greatly reduced or shut down (K Jha, manuscript in preparation, 1994) Therefore, variations in the rate of extension over time,

19 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT 1545 Table 4 Factors That Produce Variations in Rift Morphology Abundance of Volcanics Spacing of Volcanic Centers Rift Zone Width Rift Zone Depth Decompression melting Depth of melting/, m Amount of extension Thermal uplift c Temperature of mantle relative to Temperature of lithosphere d, f solidus a-e 3-D mantle flow tc'm Thickness of lithosphere a,/ Amount of extension Temperature and thickness of Low mantle viscosity Presence of melt c, n and thinning lithosphere a, f Low spreading rate Amount of extension a'c, e, g Low permeability Buoyancy of depleted Rate of extension/, n mantle residuum c, i Buoyant mantle flow h'm Mantle viscosity Underplating and Permeability volcanic infill c awhite et al [1987] bmckenzie and Bickle [1988] cwhite and McKenzie [1989] dgriffiths and Campbell [1991] emohr [1992] fhill [ 1991 ] gmohr [1983] hbuck and Su [1989] isotin and Parrnentier [1989] JCordery and Phipps Morgan [1992] CSparks et al [1992],[1993] ljha et al [1992],[1993] ink Jha, manuscript in preparation, 1994 nwhite [1992] or along different portions of the rift may affect the volume of associated with the arrival of a large mantle diapir or plumelet melt produced Variations in mantle viscosity and permeability [Pronin and Stofan, 1990; Stofan and Head, 1990; Stofan et al, will also influence upwelling and the generation of melt along a 1991, 1992; Magee Roberts et al, 1992] In addition, the rift [Buck and Su, 1989; Cordery and Phipps Morgan, 1992] origin of radially fracturedomes or novae has been interpreted Variations in any of these factors along the riff may account for as the result of early-stage uplift associated with corona variations in the amount of volcanism observed and the poor evolution [Janes et al, 1992; Squyres et al, 1992a], although correlation between volume of melt and amount of extension this hypothesis is under some debate [Pat3'i'tt amt Head, 1992, along the rift Thus, riff-related upwelling remains a viable 1993a; Parfitt et al, 1992] Where mantle plumelets have mechanism for the production of melt along the Lada rift intersected the rift, more substantial amounts of volcanism are produced than elsewhere Compared to other coronae within Active Upwelling Lada Terra, the coronae and volcanic centers along the rift are An alternative to riff-related upwelling is active upwelling o! associated with much larger amounts of volcanism (Figures! 5 relatively deep mantle plumes Previously we discussed the and 16) Excluding Quetzalpetlatl and Derceto, volcanic centers importance of preexisting riffs on the production of substantial whose relationship to the rift is unclear, the average flow field amounts of volcanism at coronae [Magee Roberts and Head, area of centers on the rift is 19 x 105 km 2, compared to 18 x 1993] Our results indicated that where corona-forming mantle 104 km 2 for centers off the rift If Quetzalpetlatl and Derceto are diapirs or "plumelets" intersected regions of lithospheric considered off-rift, the mean value for flow field area increases thinning and extension they underwent greater amounts of to 12 x 105 km 2, but is still less than that for centers on the pressure-release melting and volcanism than at coronae rift A comparison of median values is even more striking elsewhere Is it plausible that the volcanicenters along the (Figure 16) However, the concentration of volcanicenters, Lada rift are the result of mantle plumelets unrelated to the particularly coronae, along the rift (Figure 15) indicates that rifting itself? Such an origin is consistent with the fact that upwellings the area have not been randomly emplaced most of the centers are either coronae or large flow fields Furthermore, it seems unlikely that such a large number of (fluctus) that have been linked to melting and deformation upwellings could have randomly intersected the riff since it

20 1546 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT lo lo -lo -3o Inninit3-3O -5O Lavinia Planitia -5O -7O -*_78-7O 0km OO Figure 15 Map showing the distribution of coronae in the Lada Terra-Lavinia Planitia region, modified from Magee Roberts and Head [1993] Triangles represent coronae with flow fields Circles represent coronae without flow fields Additional volcanic centers with and without flow fields along the Lada rift (Mylitta, Kaiwan, Carpo, and the unnamed nova) are also shown by the same symbols Symbols are overlain on a Mercator projection of contoured Pioneer Venus altimetry at 1-km intervals The zero contour (bold) corresponds to 6051 km The approximate area of Lada Terra is shaded The Lada rift-lavinia Planitia region is highlighted by a black box Note the near total absence of coronae in Lavinia and the concentration of coronae with flow fields and other volcanic centers along the Lada- Lavinia border formed, although the time scale for deformation associated with response to downwelling under Lavinia, similar to the influx of mantle downwelling (assuming that rifting is related to mantle asthenosphere that occurs following an episode of downwelling beneath Lavinia) is unconstrained [Bindschadler delamination on Earth [eg, Bird, 1979; Kay and Mahlburg and Parmentier, 1990; Bindschadler et al, 1992] Kay, 1993] This mechanism is similar to the "flanking counterflows" associated with mantle downwelling described by Mantle Flow Related to Downwelling Zandt and Carrigan [1993] in a model for the evolution of a region of California, including the San Joaquin Valley and A third possibility is that volcanism along the Lada rift is adjacent southern Sierra Nevada mountains According to Zandt related to some form of mantle return flow that has occurred in and Carrigan, mantle downwelling resulted in downward flexure i,,,i,,,, I,,,I,, of the lithosphere that formed a sedimentary basin on the, I,,,, I,, I,,,, surface These downward forces and mantle counterflows d 'l0 106 ' 05 o ø Selu Ouetzalpetlatl Eve / Oerceto /l aiwa M yl,t, a on rift- I - off rift Figure 16 Plot of flow field area for coronae and volcanic centers along the rift and off the rift within Lada Terra (plotted in shaded area in Figure 15) The average flow field area is greater for volcanic centers along the rift than off the rift, suggesting that regional extension along the rift has enhanced pressure-release melting and the production of large-scale flow units relative to volcanic centers off rift flanking the downwelling also contributed to the tilt and uplift of the adjacent Sierra Nevada mountains The concentration of volcanic centers about the margin of Lavinia along the Lada rift (Figure 15) is what might be expected, if downwelling beneath Lavinia is axisymmetric [Bindschadler et al, 1992] and has resulted in axisymmetric return flow (Figure 13) As in the previous model, where upwelling has intersected the rift, the greatest amounts of volcanism were produced As for the model of rift-related upwelling, however, it is unclear whether upwelling of normal mantle would be hot enough to produce the scale of volcanism observed The temperature of the return flow would be at least partially dependent on the depth from which upwelling begins Coronae along the rift range in diameter from -170 to 510 km Assuming the diameter of a corona is representative of the scale of its associated mantle upwelling [Magee Roberts and Head, 1993], and that a mantle upwelling can attain a diameter through processes of entrainment [Grt'ffiths and Campbell, 1990] no greater than its depth of origin, then upwelling in the Lavinia ada area must have begun at depths at least as great as kin Given this depth range and predictions that

21 MAGEE AND HEAD: ROLE OF RIFI NG IN THE GENERATION OF MELT 1547 Venus upper mantle temperatures are elevated such that greater Evolution of the Lavinia P!anitia-Lada Terra amounts of melting may be produced from an upwelling than on Region Earth, it is plausible that the observed amounts of volcanism The model of mantle downwelling underneath Lavinia along the Lada rift could be produced in this manner To summarize, three possible mechanisms for the origin of Planitia, as described by Bindschadler et al [1992], has the volcanism along the Lada rift have been considered Riftpotential to explain not just the features of Lavinia, but as we related upwelling may be able to account for the volume of melt have attempted to show in this paper, many of the characteristics of the Lada rift and its associated volcanics as observed, given estimates for higher upper mantle temperatures on Venus than on Earth Variations in local thermal structure, well The following sequence of events is based on this model lithospheric thickness, mantle viscosity and permeability, and and is an interpretation of the possible evolution of the Lavinia rifting rate may account for the differences in the volume of Planitia-Lada Terra region (Figure 13) melt produced along the Lada rift and at other sites of rifting on 1 The development of a sustained region of mantle Venus (Table 4) The intersection of multiple, higher downwelling resulted in flexure [Zandt and Carrigan, 1993] and basin subsidence [Bindschadler et al, 1992] and the formation temperature, mantle plumelets with the Lada rift may also account for the scale of volcanism observed However, such a of the Lavinia basin Downward flexure [Solomon and Head, coincidence between rifting and presumably random upwellings 1980] and the onset of crustal shortening [Bindschadler et al, seems improbable given the number and concentration of 1992] related to downwelling initiated compressional centers along the rift Mantle return flow produced in response deformation within the interior of the basin (Figure 13b) to downwelling beneath Lavinia may account for both the volme of melt observed and the distribution of volcanic cen- 2 In addition, the onset of downwelling produced tensional radial stresses about the periphery of the basin These stresses may have been sufficient to produce the extension observed along the Lada rift Alternatively, subsequent dike emplacement may have enhanced the observed amount of extension The extension was most developed along the highland border of Lada due possibly to its closer proximity to the center of downwelling and/or to the lower strength of the ters along the rift Rift-related upwelling and/or mantle counterflow related to downwelling seem the most plausible models for the origin of volcanism along the Lada rift The origins of the two large volcanic centers, Derceto and Quetzaltpetlatl, which are not located directly on the rift axis, are not clear The morphology and large scale of Quetzalpetlatl suggests that is most likely the result of an upwelling mantle thicker highland crust (Figures 13b and 13c) In addition, diapir or plumelet unrelated to rifting Derceto is much smaller in scale and more closely situated to the rift than Quetzalpetlatl downwelling-related extension along the western border of Lavinia may have been overprinted by tectonic activity associated with the formation of Dione Regio and Themis It may have formed in a manner similar to that of volcanic Regio centers along the rift, which would imply that mantle upwelling 3 Upwelling mantle, perhaps in some form of axisymmetric was not everywhere focused along the rift axis return flow associated with the downwelling beneath Lavinia, How does the origin of volcanism along the Lada rift or upwelling related to peripheral rifting about the basin, compare to corona chains such as Parga Chasma? Both are brought hotter material up from depth that subsequently proposed to be sites of extension along which mantle underwent substantial amounts of pressure-release melting upwelling has produced coronae and other volcanic centers If coronae chains such as that along Parga Chasma are also related along the rift This resulted in the formation of multiple volcanic centers along the rift axis and large floods of lava that to rifting and passive upwelling, what are the causes for the poured downslope into Lavinia (Figure 13c) differences in spacing and abundance of centers relative to the 4 Continued extension has deformed most of the volcanic Lada rift? The coronae along Parga are much greater in number centers along the rift, except Mylitta Fluctus, which remains with a smaller average spacing, diameter and flow field area relatively undeformed Continued downwelling and crustal than coronae along the Lada rift (Table 2) Variations in the shortening produced linear ridge and fracture belts within along-axis production of melt in a rifting environment ("three- Lavinia (Figure 13c) With time, continued downwelling is dimensional flow")are thought to be the result of focused predicted to produce a block of highly deformed, thickened crust upwellings produced when buoyant mantle flow dominates flow or tessera elevated above the original basin [Bindschadler and due to plate spreading; this occurs under conditions of low Parmentier, 1990; Bindschadler et al, 1992] spreading rate and mantle viscosity [Jha et al, 1992, 1993; An unresolved issue is the origin of the bulk of dark plains Sparks et al, 1992, 1993] In the absence of an within Lavinia The oldest plains in Lavinia, those associated asthenosphere, as on Venus, buoyant flow produces an alongwith local clusters of shields, may predate basin formation axis wavelength that scales with the depth of melting [Jha et The emplacement of younger plains appears to have been al, 1993] Differences in spacing and scale of volcanics interleaved with interior compressional deformation and the between Parga Chasma and the Lada rift may be due to differproduction of ridge and fracture belts [$quyres et al, 1992b] ences in the depth of melting such that volcanic centers along Sources for the younger radar-dark plains have not been Parga are associated with shallower depths of melting identified Volcanic centers along the rift are sources for the Variations in mantle temperature, lithospheric thickness, the total amount of extension along a rift, and rifting rate may account for differences in the depth of melting at different generally radar-bright digitate flow fields that are superposed on the dark background plains and embay local ridge and fracture belts Thus, the digitate flow fields appear to be the rifting sites Differences in depths of melting and the degree to youngest plains units within Lavinia Perhaps the earliest which three-dimensional buoyant flow is developed may account for the differences in the evolution of volcanism along the Lada rift and coronae chains such as Parga Chasma stage of volcanism associated with upwelling along the rift produced vast sheets of lava that flooded and filled the basin This is supported by the presence of an extensive flow field that

22 1548 MAGEE AND HEAD: ROLE OF RIb-TING IN THE GENERATION OF MELT is partially overlain by Mylitta Fluctus [Magee Roberts et al, 1992] Based on mapped flow directions, this flow field appears to have emanated from a region (now obscured) along the rift to the west of Mylitta and appears to have exceeded Mylitta in total area It is highly degraded with individual, digitate flows barely discernible In this scenario, volcanic centers or fissures along the rift may have been the sources for much of the younger background plains of Lavinia Volcanic activity may have subsequently decreased in scale and become localized at discrete centers along the rift However, it seems unlikely that the vast amounts of lava needed to resurface most of Lavinia could have been produced and erupted from sources within the rift alone An alternative, and more likely, possibility is that the majority of the plains were erupted from now buried fissures and other vents within Lavinia First-order tests for the model of the evolution of the Lavinia-Lada region described above should be possible with the acquisition of high-resolution gravity data by Magellan In the current, nearly circular orbit of the spacecraft, gravity data based on line-of-sight acceleration have a resolution of km in the high-latitude regions of Venus and may be just sufficient to resolve the Lada rift A negative gravity anomaly centered over eastern Lavinia was observed in the data obtained by Pioneer Venus and interpreted to indicate that basin subsidence is compensated by a mass anomaly at large depths [Bindschculler et al, 1992] Such an anomaly should be apparent in the Magellan data A gravity signature of the Dada rift was not detected in the Pioneer Venus data Unless it is swamped by the signature of the Lada highland, a signature of the rift may be discernible in the Magellan gravity data and provide information regarding the manner by which rift topography is compensated and the depth of associated mantle upwelling The model described above also has implications for the possible origin and evolution of other lowlands on Venus, such as Atalanta Planitia, that are also nearly circular in planform and have been related to mantle downwelling [Bindschadler et al, 1992] If an examination of the region about Atalanta reveals evidence of rifting and associated volcanism, then it may be possible to compare the evolutionary stages of downwelling for the two areas and to evaluate various parameters such as the presence of a nearby highland (as a zone of weaker crust) in the development of rifting and volcanism as described for Lada Terra Factors Controlling Variations in Scale and Morphology of Volcanism continued volcanism (primarily interior flooding) and degradation of the annulus [Pronin and Stofan, 1990; Stofan and Head, 1990; Stofan et al, 1991, 1992; Janes et al, 1992; Squyres et al, 1992a] According to this model, features like Carpo and the unnamed nova that are dominated by radial fractures would represent the early stages of corona formation Features characterized by annuli and radial fractures, such as Tamfana and Selu, would represent the later stages However, many of the volcanic centers have morphologies inconsistent with the published evolutionary scheme of coronae For example, Eve, Eithinoha and Sarpanitum all lack well-defined radial fracture systems that are not related to continued extension along the rift To be consistent with the standard model of corona evolution, such features would have to have been partly or wholly obscured by subsequent volcanism In addition, no early stage nova-like radial fractures nor any pronounced uplift are observed at either Kaiwan or Mylitta Fluctus As mentioned earlier, the scale of volcanism varies significantly among the different volcanic centers Some centers are dominated by extensive flow fields, some have more modest flow fields that formed at different times relative to the corona annulus, and others lack flow fields altogether Clearly, there are coronae along the rift, and some (for example, Selu) appear to have evolved in a manner consistent with published models However, there is no compelling evidence that each center along the rift is evolving into a corona or that each corona has followed exactly the same evolutionary path Variations in the stage of corona evolution are less likely to account for the range of volcanic morphologies observed along the rift than variations in the processes of mantle upwelling, melt formation, and magma ascent Clearly, the standard model of corona formation is consistent with the apparent evolution of many such features on Venus [eg, Stofan et al, 1992] However, based on our examination of volcanic features associated with the Lada rift, it is apparent that coronae formation has occurred in multiple, complex combinations of the processes of mantle upwelling, melt formation, and magma ascent, and is not adequately represented in all cases by any single evolutionary sequence An alternative possibility is that variations in the scale of mantle upwelling associated with each center have produced the variations in volcanic morphology along the rift such that the largest diapirs or plumelets have produced the greatest amounts of partial melt and volcanism and possibly the greatest mnounts of uplift Previously, we examined this hypothesis for the population of coronae as a whole and found that there is very little correlation between the size of a corona (taken as an indicator of the scale of upwelling) and its associated flow field Volcanic centers along the Lada rift have a variety of [Magee Roberts and Head, 1993] The greatest factor in the morphologies, ranging from coronae to novae and large flow scale of output for coronae as a whole is the presence of locally fields There are variations among the various types of stretched and thinned lithosphere that allows mantle volcanic centers as well For example, some coronae along the upwellings to undergo greater amounts of pressure-release rift possess flow fields (such as Eve) and/or radial structures melting than elsewhere Could variations in the scale of (Selu, Tamfana), and others do not (Sarpanitum) There are at least three possible reasons for the variation in morphology and volcanic output along the rift upwelling be more significant in modifying the scale of volcanism of centers actually within a region of thinned lithosphere? Among coronae along the Lada rift, the largest One explanation is that each center represents a different flow fields are generally associated with the larger coronae stage in the evolution of a corona or mantle plumelet Corona evolution has been modeled to include the following stages: (1) uplift due to the impingement of a mantle diapir against the (eg, Eve and Selu; Figure 17) However, the largest corona, Eithinoha, does not possess the largest flow field, and the correlation coefficient (p) between corona diameter and flow lithosphere, radial fracturing, and volcanism; (2) subsidence field area is only 051 Eithinoha is possibly the oldest corona and annulus formation due to gravitational relaxation of the elevated topography and/or magma withdrawal; and (3) along the rift It may have formed prior to rifting, and its flow field may have degraded over time such that only a portion is

23 - - _ - MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT E 10 I , _, : ' _o - ( a ) coronae on rift ß Selu ß Eve release melting experienced by an upwelling and the ability of the melt to rise to the surface before solidifying These factors may explain the variations in the amount of uplift at volcanic centers along the rift as well as the relative importance of radial fracturing (both as a result of uplift and dike emplacement), annulus formation (the greater uplifts having experienced the most relaxation), and scale of volcanic output Comparison With Formation of Terrestrial Flood Basalts 00 I 0 ø E e o oo 1 o ø ß Eithinoha Sarpanitum Tamfana ''' 1''''1'O''1''''1 I''''1 I (b) coronae off rift ß Derceto - ß corona diameter (km) Quetzalpetlatl,= =,,=, =,=,,,, rona diameter(km) 900 9OO Figure 17 Plot of flow field area versus diameter for coronae (a) along the Lada rift and (b) off the rift within Lada Terra Jord and Kamui-Huci are not included as they may possess flow fields that have either degraded or been superposed by flows of an adjacent center and cannot be accurately examined (see text for discussion) The correlation between flow field area and scale of upwelling is generally poor (9-051) for coronae along the rift For coronae off the rift, Factors other than the scale of upwelling appear to be the dominant control on the variation in the scale of volcanism associated with coronae along the rift still visible in the radar data Even so, it appears that factors other than the scale of upwelling are required to explain the variations in volcanism along the rift As discussed earlier, additional factors in the variation of volcanic style along the rift include variations in local thermal structure, crustal and lithospheric thickness, and total amount of extension, manfie viscosity and permeability, and changes in rifting rate over time (Table 4) [eg, Jha et al, 1992, 1993, and K Jha, manuscript in preparation, 1994] Along the trend of the rift are blocks of tessera and variations in local topography that might reflect variations in local crustal thickness and/or the amount of extension Variations in the Flow fields associated with volcanic centers along the Lada rift range in area from 3370 km 2 to over 6 x 105 km 2 The four largest range from 3 to 6 x 105 km 2 (Table 1) These flow fields are similar in scale to many terrestrial flood basalts, which range from-1 x 105 to 2 x 106 km 2 [eg, Hooper, 1990] Terrestrial flood basalts are thought to form as a result of pressure-release melting in the head of a mantle plume [eg, Richards et al, 1989; Campbell and Gr!ffiths, 1990] Similarly, large-scale volcanism characteristic of "great flows" or fiuctus and associated with coronae on Venus has been related to pressure-release melting in an upwelling mantle diapir or plumelet [Magee Roberts et al, 1992; Magee Roberts and Head, 1993] Terrestrial flood basalts are frequently correlated with episodes of continental rifting and breakup, although the relative timing of rifting and volcanism may vary White and McKenzie [1989] proposed that the formation of flood basalts requires the upwelling of anomalously hot mantle (ie, a mantle plume) under conditions of preexisting lithospheric stretching and thinning Other researchers have observed that many flood basalts actually erupt prior to major episodes of rifting (eg, the Deccan Traps, the Karoo basalts) or are not associated with continental breakup or large amounts of extension (eg, the Columbia River basalts, the Siberian Traps) and proposed that lithospheric extension is not a prerequisite for the formation of flood basalts [Richards et al, 1989; Campbell and Griffiths, 1990; Hooper, 1990] Recently, White [1992] and Saunders et al [1992] noted that although complete continental breakup and seafloor spreading may not be associated with all flood basalt provinces, some of these provinces, such as the Columbia River, appear to have been associated with limited amounts of lithospheric thinning or stretching in its initial stages which may have been sufficient to produce voluminous amounts of decompression melting within a mantle plume Thus, it is also likely that the timing of rifting and the emplacement of flood basalts would appear to vary, depending on the relative stage of continental breakup In order to rapidly generate huge volumes of melt, the upwelling mantle or plume head must decompress to depths of a few tens of kilometers, which according to White [1992] is not possible without lithospheric thinning and stretching Alternatively, differences in subcontinental lithospheric thicknesses and temperatures may account for variations in the timing of rifting and the eraplacement of terrestrial flood basalts [Hill, 1991] According to this model, when the lithosphere is relatively cool (-500øC) and thick (-200 km), amount of extension and in local crustal thickness may have areally extensive basalts are erupted my prior to the affected the depth to which mantle upwellings could rise along onset of rifting, depending on the time needed for heat the rift Variations in these parameters, as well as in local conduction to weaken the uppermost mantle and lowermost thermal structure, mantle viscosity and permeability, and time- crust When the lithosphere is relatively hot (-700øC) and thin dependent changes in extension rate (which cannot be evaluated (-120 km), heat conduction is rapid and the initiation of in the current data), would have affected the amount of pressure- extension may predate the eruption of voluminous basalts

24 155o MAGEE AND HEAD: ROLE OF RIbWING IN THE GENERATION OF MELT The association of terrestrial flood basalts with the impingement of plume heads at the base of the lithosphere was recently challenged by Anderson et al [1992] on the basis of the absence of plume heads under flood basalt provinces (particularly regions too young to have significantly cooled) as observed in high-resolution three-dimensional seismic tomography of the upper mantle Anderson et al [1992] proposed that extensive basaltic volcanism is produced when lithospheric extension occurs over particularly hot regions of the mantle; mantle plumes are not necessary This has relevance to the interpretation of the origin of large-volume flow units associated with the Lada rift on Venus, as passive upwelling of mantle related to rifting and/or mantle return flow related to downwelling (not large-scale mantle plumes) are thought to be the cause for melt generation Large-scale flow units may have been eraplaced along the 'Lada rift as the lithosphere was extended over a relatively hot region of Venus mantle If the sun'ounding mantle is very hot, perhaps the "coldspot" associated with the downwelling under Lavinia is cold only in a relative sense Although it is not yet known whether preexisting extension is required for the formation of all large-scale flow units on Venus, limited extension (or a tensional stress regime) appears to have been active prior to the eruption of large flow fields along the Lada rift Preexisting conditions of lithospheric extension were also shown to be associated with the eruption of large-scale flow units at coronae on Venus [Magee Roberts and Head, 1993] Perhaps higher mantle temperatures and possibly thinner lithospheric thicknesses [eg, Stevenson et al, 1983; Solomon and Head, 1991] on Venus cause regions of mantle plume activity to undergo rapid thinning and extension, prior to the formation of large-scale flow units as described by Hill [1991] for the Earth Or, perhaps, greater lithospheric thicknesses [eg, Turcotte, 1993] on Venus inhibit the formation of flood-type eruptions except where mantle upwellings happen to intersect (or are associated with) regions of rifted and thinned lithosphere Continued analysis of largescale flow units on Venus is required to answer questions regarding the formation of these possible analogs to terrestrial flood basalts and their association with zones of extension (K P Magee and J W Head, manuscript in preparation, 1994) Summary and Conclusions Senske et al, 1992; Solomon et al, 1992] but may be linked to peripheral extension associated with mantle downwelling underneath Lavinia Planitia [Bindschadler and Parmentier, 1990; Bindschadler et al, 1992] Relatively shallow mantle upwelling due to rifting and/or counterflow associated with downwelling beneath Lavinia may account for both the volume of melt observed and the distribution of volcanic centers along the rift Pressure-release melting of the upwelling mantle was enhanced where it intersected the stretched and thinned regions of the Lada rift Variations in the scale and morphology of volcanic centers along the rift are primarily the result of variations in the local crustal and lithospheric structure and thickness The presence of numerous coronae and other volcanic centers is characteristic of the Lada rift and features known as corona chains [Senske, 1992; Senske and Head, 1992; Stofan et al, 1993] Both structures are thought to be related to mantle upwelling along linear zones of extension [Stqfan et al, 1993] Corona chains are comprised of a greater number of coronae that are also smaller in average diameter, more closely spaced, and associated with smaller flow fields than those located within the 'Lada rift Variations in the scale and distribution of volcanic centers along a rift may be due to differences in the depth of melting and the degree to which three-dimensional buoyant mantle flow is developed [Jha et al, 1992, 1993; Sparks et al, 1992, 1993] Although the origin of both extension and volcanism along the Lada rift may be related to processes associated with mantle downwelling underneath Lavinia Planitia, rifting has clearly enhanced the production of pressure-release melting as observed previously in our study of volcanism associated with coronae on Venus [Magee Roberts and Head, 1993] Whether or not the presence of thinned and rifted lithosphere is necessary for the formation of all large-scale flow units on Venus is the subject of further research (K P Magee and J W Head, manuscript in preparation, 1994) Acknowledgments We gratefully acknowledge grant NAGW-1873 from NASA (to J W H)and support from the William F Marlax Memorial Foundation Thanks axe extended to Maxc Paxmentier and Liz Parfitt for helpful discussions of this work Peter Hooper and Duane Bindschadler provided helpful reviews of the manuscript Thanks are also extended to Peter Neivert for photographic assistance References Rifting of the 'Lada Terra highland occurs over a distance in excess of 6000 km along the border of the adjacent lowland, Anderson, D L, Y-5 Zhang, and T Tanimoto, Plume heads, Lavinia Planitia, and is characterized by a series of continental lithosphere, flood basalts and tomography, in Magmatism and the Causes of Continental Break-up, edited by BC Storey, T discontinuous topographic lows km deep In contrast Alabaster, and R I Pankhurst, Geol Soc Spec Publ London, 68, to other rifts on Venus, such as Devana Chasma at Beta Regio 404 pp, 1992 [Senske, 1992; Senske and Head, 1992; Senske et al, 1992], Baer, G, G Schubert, D L Bindschadler, and E R Stofan, Spatial and deformation is not restricted to a single trough bound by steep temporal relations between coronae and extensional belts, northern scarps, nor do multiple rift arms radiate from a major central volcanic edifice Broader zones of fracturing km wide IaAa Terra, Venus, J Geophys Res, 99(E4), , 1994 Bindschadler, D I and E M Paxmentier, Mantle flow tectonics: The flank the deepest portions of the rift Multiple volcanic influence of a ductile lower crest and implications for the formation centers, including large coronae, radially fractured edifices of topographic uplands on Venus, J Geophys Res, 95 (B 13), known as novae, and large flow fields similar in scale to 21,329-21,344, 1990 terrestrial flood basalts are located along the rift with an av- B indschadler, D L, G Schubert, and W M Kaula, Coldspots and hotspots: Global tectonics and mantle dynamics of Venus, J erage spacing of 630 km Volcanism postdates the onset of Geophys Res, 97(E8), 13,495-13,532, 1992 extension, although continued rifting has deformed many of the Bird, P, Continental delamination and the Colorado Plateau, J volcanic centers along the rift Geophys Res, 84(B13), , 1979 In contrast to other Venus rifts, the origin of rifting is not Buck, W R and W Su, Focused mantle upwelling below mid-ocean the result of uplift and extension associated with the presence of a large mantle plume [eg, Grimm and Phillips, 1992; ridges due to feedback between viscosity and melting, Geophys Res Lett, 16, , 1989

25 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT 1551 Campbell, D B, D A Senske, J W Head, A A Hine, and P C Fisher, Venus southern hemisphere: Geologic character and age of terrains in the Themis-Alpha-Lada region, Science, 251, , 1991 Campbell, I H and R W Griffiths, Implications of mantle plume structure for the evolution of flood basalts, Earth Planet Sci Lett, 99, 79-83, 1990 Cordery, M J and J Phipps Morgan, Melting and mantle flow beneath amid-ocean spreading center, Earth Planet Sci Lett, 111, , 1992 Crisp, J A, Rates of mag na emplacement and volcanic output, J Volcanol Geotherm Res, 20, , 1984 Dieterich, J H, Growth and persistence of Hawaiian volcanic rift zones, J Geophys Res, 93(B5), , 1988 Dvorak, J J, / T Okamura, T T English, R Y Koyanagi, J S Nakata, M K Sako, W T Tanigawa, and K M Yamashita, Mechanical response of the south flank of Kilauea Volcano, Hawaii, to intrusive events along the rift systems, Tectonophysics, 124, , 1986 Ford, P G and G H Pettengill, Venus topography and kiloineter-scale slopes, J Geophys Res, 97(E8), 13,103-13,114, 1992 Gibson, I L, N S Madhurendra, and W F Fahrig, The geochemistry of the MacKenzie dyke swarm, Canada, Geol Assoc Can Spec Pap 34, , 1987 Griffiths, R W and I H Campbell, Stirring and structure in mantle starting plumes, Earth Planet Sci Lett, 99, 66-78, 1990 Griffiths, R W and I H Campbell, Interaction of nantle plume heads with the Earth's surface and onset of s nall-scale convection, J Geophys Res, 96(Bll), 18,295-18,310, 1991 Grimm, R E and R J Phillips, Tectonics of Lakshmi Planum, Venus: tests for Magellan, Geophys Res Lett, 17, , 1990 Grimm, R E and R J Phillips, Anatomy of a Venusian hot spot: geology, gravity and mantle dynanfics of Eistla Regio, J Geophys Res, 97(E10), 16,035-16,054, 1992 Hamilton, V E and E R Stofan, Morphology and xnodels lution of Eastern Hecate Chasma, Venus (abstract), Datar Planet Sci, XXIV, , 1993 Head, J W, L S Crumpier, J C Aubele, J E Guest, and R S Saunders, Venus volcanism: Classification of volcanic features and structures, associations, and global distribution from Magellan data, J Geophys Res, 97(E8), 13,153-13,197, 1992 Hill, R I, Starting plumes and continental break-up, Earth Planet Sci Lett, 104, , 1991 Hooper, P R, The timing of crustal exteasion and the eruption of continental flood basalts, Nature, 345, , 1990 Janes, D M, S W Squyres, D L Bindschadler, G Baer, G Schubert, V I Sharpton, and E R Stofan, Geophysical models for the formation and evolution of coronae on Venus, J Geophys Res, 97(E10), 16,055-16,067, 1992 Jha, K, E M Parmentier, and J Phipps Morgan, Mantle flow beneath spreading buoyancy (abstract), Eos Trans AGU, 73, Spring Meeting Suppl, Magee Roberts, K, J E Guest, J W Head, and M G Lancaster, Mylitta Fluctus, Venus: Rift-related, centralized volcanism and the emplacement of large-volume flow units, J Geophys Res, 97(E10), 15,991-16,015, 1992 McKenzie, D and M J Bickle, The volume and composition of melt generated by extension of the lithosphere, J Petrol, 29, , 1988 Mohr, P, Ethiopian flood basalt province, Nature, 303, , 1983 Mohr, P, Nature of the crust beneath magmatically active continental rifts, Tectonophysics, 213, , 1992 Parfitt, E A and J W Head, Radial fracture systeins on Venus: Conditions of for nation (abstract), Lunar Planet Sci, XXIII, 75-76, 1992 Parfitt, E A and J W Head, Formation and evolution of radial fracture systems on Venus (abstract), Lunar Planet Sci, XXIV, , 1993a Parfitt, E A and J W Head, Buffered and unbuffered dike e nplacement on Earth and Venus: Implications for magma reservoir size, depth, and rate of magma replenishment, Earth Moon Planets, 61, , 1993b Parfitt, E A, L Wilson, and J W Head, The origins of radial fracture systems and associated large lava flows on Venus (abstract), in International Colloquium on Venus, pp 83-84, Lunar and Planetary Institute, Houston, Tex, 1992 Phillips, R J, Convection-driven tectonics on Venus, J Geophys Res, 95(B2), 1301-!316, 1990 Phillips, R J and M C Malin, The interior of Venus and tectonic implications, in Venus, edited by D M Hunten, L Colin, T M Donahue, and V I Moroz, pp , University of Arizona Press, Tucson, 1983 Phillips, R J, R E Arvidson, J M Boyce, D B Campbell, J E Guest, G G Schaber, and L A Soderblom, Impact craters on Venus: Initial analysis from Magellan, Science, 252, , 1991 for the evo- Pronin, / A and E R Stofan, Coronae on Venus: Morphology, classification, and distribution, Icarus, 87, , 1990 Richards, M A, R A Duncan, and V E Courtillot, Flood basalts and hot-spot tracks: Plume heads and tails, Science, 246, , 1989 Roberts, IC M, J W Head, M G Lancaster, and J E Guest, Volcanism and rifting along the northern edge of Lada Terra, Venus (abstract), LunarPlanet Sci, XXIII, 1! , 1992 Saunders, AD, M Storey, R W Kent, and M J Norry, Consequences of plume-lithosphere interactions, in Magmatism and the Causes of Continental Break-up, edited by BC Storey, T Alabaster, and R J Pankhurst, Geol Soc Spec PubL London, 68, 404 pp, 1992 Senske, D A, Zones of extension on Venus: Characteristics, distribution, examination of the style of rifting and evaluation of a plate flexure model for Devana Chasma at Beta Regio, PhD thesis, pp , Brown Univ, Providence, R I, 1992 Zones of extension and rifting on Venus: centers due to mantle depletion and melt retention Senske, D A and J W Head, Characteristics and distribution (abstract), Lunar Planet Sci, XX!11, 291, , 1992 Jha, K, E M Parmentier, and J Phipps Morgan, Episodic three-dimen- Senske, D A et al, Geology and tectonics of the Themis Regiosignal mantle upwelling and crustal production beneath a spreading Lavinia Planitia-Alpha Regio-Lada Terra area, Venus: Results from center (abstract), Eos Trans AGU, 74, Spring Meeting Suppl, 304, Arecibo image data, Earth Moon Planets, 55, , Senske, D A, G G Schaber, and F R Stofan, Regional topographic Kay, R W and S Mahlburg Kay, Delamination and delamination mag- rises on Venus: Geology of Western Eistla Regio and comparison to matism, Tectonophysics, 219, , 1993 Beta Regio and Atla Regio, J Geophys Res, 97(E8), 13,395- Keddie, S T, Preliminary analysis of Dione Regio, Venus: The final 13,420, 1992 Magellan regional imaging gap (abstract), Lunar Planet Sci, XXIV, Solomon, SC and J W Head, Lunar mascon basins: Lava filling, tec , 1993 tonics, and evolution of the lithosphere, Rev Geophys, 18, 107- Lancaster, M G, J E Guest, IC M Roberts, and J W Head, "Great" 141, 1980 lava flow fields on Venus (abstract), Lunar Planet Sci, XXIII, 753- Solomon, SC and J W Head, Fundamental issues in the geology and 754, 1992 geophysics of Venus, Science, 252, , 1991 Latin, D and N White, Generating melt during lithospheric extension: Solomon, SC, J W Head, W M Kaula, D McKenzie, B Parsons, R Pure shear vs simple shear, Geology, 18, , 1990 J Phillips, G Schubert, and M Talwani, Venus tectonics: Initial Magee Roberts, K and J W Head, Large-scale volcanism associated analysis from Magellan, Science, 252, , 1991 with coronae on Venus: Implications for formation and evolution, Solomon, SC eta!, Venus tectonics: an overview of Magellan Geophys Res Lett, 20, , 1993 observations, J Geophys Res, 97(E8), 13,199-13,255, 1992

26 1552 MAGEE AND HEAD: ROLE OF RIFTING IN THE GENERATION OF MELT Sotin, C and E M Parmentier, Dynanfical consequences of composi- Implications for origin and relation to mantle processes, J tional and thermal density stratification beneath spreading centers, Geophys Res, 97(E8), 13,347-13,378, 1992 Geophys Res Lett, 16, , 1989 Stofan, R, V E Hamilton, and K Cotugno, Parga and Hecate Sotin, C, D P Senske, J W Head, and M Parmentier, Terrestrial Chasmata, Venus: Structure, volcanism and models of formation spreading centers under Venus conditions: Evaluation of a crustal (abstract), LunarPlanet Sci, XXIV, , 1993 spreading model for Western Aphrodite Terra, Earth Planet Sci Swanson, D A, W P Duffield, and R S Fiske, Displacement of the Lett, 95, , 1989 south flank of Kilauea Volcano: The result of forceful intrusion of Sparks, D W, E M Parmentier, and J Phipps Morgan, 3-D numerical magma into the rift zones, US Geol Surv Prof Pap, 963, 1976 experiments on buoyant flow beneath segmented spreading centers Turcotte, D, Is there uniformitarian catastrophic tectonics Venus (abstract), Eos Trans AGU, 73, Spring Meeting Suppl, 291, 1992 (abstract), LunarPlanet Sci, XXIV, , 1993 Sparks, D W, E M Parmentier, and IC Jha, The effect of melt migra- White, R S, Magmatism during and after continental break-up, in tion on 3-D mantle flow beneath spreading centers: relativeffects Magmatism and the Causes of Continental Break-up, edited by BC of thermal, compositional, and melt-retention buoyancy (abstract), Storey, T Alabaster_, and R I Pankhurat, Geol Soc Spec Publ Eos Trans AGU, 74, Spring Meeting Suppl, 304, 1993 London, 68, 404 pp, 1992 Squyres, S W, D M Janes, G Baer, D L Bindscharier, G Schubert, White, R and D McKenzie, Magmatism at rift zones: the generation of V L Sharpton, and E R Stofan, The morphology and evolution of coronae on Venus, J Geophys Res, 97(E8), 13,611-13,634, 1992a Squyres, S W, D G Jankowski, M Simons, SC Solomon, B H volcanic continental margins and flood basalts, J Geophys Res, 94(B6), , 1989 White, R S, G D Spence, S R Fowler, D P McKenzie, G K Westbrook, and P N Bowen, Magmatism at rifted continental Hager, and G McGill, Plains tectonism on Venus: The margins, Nature, 330, , 1987 deformation belts of Lavinia Planitia, J Geophys Res, 97(E8), Zandt, G and C R Camgan, Small-scale convective instability and 13,579-13,599, 1992b upper mantle viscosity under California, Science, 261, , Stevenson, D J, T Spohn, and G Schubert, Magnetisdn and thermal 1993 evolution of the terrestrial planets, Icarus, 54, , 1983 Stofan, E R and J W Head, Coronae of Mne nosyne Regio: Morphology and origin, Icarus, 83, , 1990 J W Head, Depart nent of Geological Sciences, Brown University, Stofan, R, D I B indschadler, J W Head, and M Parmentier, Providence, RI Coronae structures on Venus: Models of origin, J Geophys Res, IC P Magee, 1729 Raymond Hill Road #4, South Pasadena, CA 96(E4), 20,933-20,946, Stofan, R, V L Sharpton, G Schubert, G Baer, D L Bindschadler, D M Janes, and S W Squyres, Global distribution and (Received November 16, 1993; revised August 8, 1994; characteristics of coronae and related features on Venus: accepted September 6, 1994)