As David Arthur comments, muchof the evidence is under the sea. Spencer Wells, in "The Journey of Man: a genetic odyssey" argued in 2002 that the speciation of H. sapiens occurred largely along the coastal plain which was exposed around 100,000 years ago from Africa almost to Australia. In my book I take his argument to suggest that the evolving dog accompanied us in an extended phenotype and the complex evolved the anatomy to speak in words. H sapiens + dog could then turn north (marshalled by the geography of the Indian ghatts and the Himalayas), confront the Neandertal in the Middle East/West Asia, out-comoete them and the rest is Cro-Magnon history. Please see www.compositeconversationalist.com

Is this entirely surprising? Modern humans tend to live along waterways either natural rivers and streams or coastal areas.

I'm not entirely sure that this indicates humans evolved in a watery environment as opposed to a terrestrial.. The savanna is where humans had generally been thought to evolve from.

Extract from the following site below: http://www.aquaticape.org/genprobs.html

An example is the assertion, echoed by almost all AAT/H versions, that paleoanthropologists claim that early hominids lived on "the arid savannah", which AAT/H proponents typically -- and inaccurately -- characterize as a waterless, treeless plain. The AAT/H is then said to be the only sensible theory, since early hominids were probably not well adapted for surviving on a treeless, waterless arid grassland. But paleoanthropologists don't make this claim; they actually refer -- quite clearly -- to an early hominid habitat consisting of mixed vegetation: trees, bushes, and grass (and some say the earliest hominid habitats were likely denser forests).

Definition of savanna:-

From 1986 The Savannahs: Biogeography and Geobotany, by Monica M. Cole (Department of Geography, Royal Holloway and Bedford New College, University of London). Academic Press: London.
pg. 4:
"An understanding of the problems that have beset the formulation of an agreed definition of savannas requires consideration of the origin, historical and present use of the word."

pp. 4-5:
"The term savanna is believed to originate from an Amerindian word which, in a work on the Indies published in 1535, was used by Oviedo y Valdes to describe 'land which is without trees but with much grass either tall or short'. Subsequently Grisebach (1872), Drude (1890) and Schimper (1903) extended its use to include grasslands with trees and thereafter the term was used to describe the mixed trees and grass types of vegetation found in all tropical latitudes."

pg. 5:
In central and southern Africa savanna is used for open deciduous woodlands, including those locally known as miombo, that are composed of fairly tall, mesophyllous trees and a well defined grass stratum, for parklike vegetation comprising grasslands studded with microphyllous trees of low to medium height, for grasslands with scattered clumps of trees or bushes, for treeless grasslands of tall perennial mesophytic grasses and of short annual grasses mixed with perennial grasses with narrow rolled leaves, and for open forms of vegetation composed of scattered low growing microphyllous trees and shrubs and a ground layer of perennial and annual grasses. The term bushveld is used locally for the parklike forms of savanna which together with the low tree and shrub forms are regarded as the most typical savannas.