Southeast of the ziggurat terrace was a huge and heavily fortified building called Gi-par-u. It was dedicated to Ningal, the consort of the moon god Nanna, but also contained several minor shrines that were incorporated during its long and complicated history. It has been identified as the residence of the entu, the king's daughter who was appointed as a high priestess. More precisely, the entu was the earthly wife of the god Nanna; she was the highest-ranking available woman and thus perfect for the god. Sargon's daughter (Naram-Sin's aunt) Enheduanna is the first entu that we know of, and the giparu likely dates to his reign.
The entu tradition begun by Naram-Sin endured for centuries. Even usurpers allowed the incumbent entu to die naturally before installing a new one. That the giparu was the home of the entu is indicated by the burial of entus at the giparu itself, and the resultant accumulation of many dead entus over time. The giparu is adjacent to the ziggurat, atop which was surely a shrine to Enlil and which was thus his earthly home. However, Enlil already had a wife (the divine Ninlil) and thus the earthly entu occupied a separate house but still resided in the complex of her divine husband's home.
Lloyd -- Woolley observed three strange aspects: the fabric of the tower was sun-dried brickwork, reinforced with thick layers of woven reeds at intervals of six or eight courses. Penetrating into its core, he found that bricks and mortar alike were hardened and discolored by a fire, a phenomenmon he attributed to the saturation of the structure by damp (perhaps while temrporarily in ruins) and the consequent occurence of internal combusting in the decaying vegetable matter. Conversely he was surprised by the multiplication of 'weeper-holes' penetrated the baked-brick outer shell of the lowest stage; and he wondered whether the terrace above acould have been planted with trees which would have required irrigation. He had in fact found carbonized tree trunks near the base of the tower. Finally he discovered a slight outward deviation in the line of the main façcades at pavement level. This he was tempted to comapre with the subtle distortion of perspective contrived by the builders of Grek temples fifteen centuries later. Remembering the grat weight of the structure and the pliable nature of mud brick, other have mistrusted this explanation.
Woolley himself says the wall around the temenos had beneath the plaster at regular intervals of two feet thesmall rounded heads of clay 'nails' driven into the mud mortar betwen the brick courses -- so-called foundation cones and on the nail's stem was the inscription 'For Nannar the strong bull of Heaven, most glorious son of Enlil, his king, has Ur-Nammu the mighty man, King of Ur, built his temple, E-temen-ni-il'.
When first we started the work of drawing out the plan and elevations of the Ziggurat we were puzzled to find that the different measurements never seemed to agree; then it was discovered that in the whole building there is not a single straight line, and that what we had assumed to be such were in fact carefully calculated curves. The walls not only slope inwards, but the line from top to bottom is slightly convex; on the ground plan the wall line from corner to corner of the building has a distinct outward bend, so that sighting along it one can see only as far as the centre; the architect has aimed at an optical illusion which the Greek builders of the Parthenon at Athens were to achieve many centuries afterwards, the curves being so slight as not to be apparent, yet enough to give the eye an appearance of strength where a straight line might by contrast with the mass behind it have seemed incurved and weak. The employment of such a device does great credit to the builders of the twenty-second century before Christ. Woolley 1954, p 131
No one looking at the Ziggurat can fail to notice the tall and narrow slits which at regular intervals and in rows one above another pierce the brickwork of the walls; they run clean through the burnt-brick casing and deep into the mud brick of the core, where they are loosely filled with broken pottery. These are 'weeper-holes' intended to drain the interior, a necessary precaution, for with damp the mud brick would swell and make the outer walls bulge if it did not burst them altogether. Woolley 1954, p 132
In the doorway of a room of late date lying against the back wall of the tower we found a great diorite hinge-stone bearing an inscription of Nabonidus in which he refers to his repairs of the building and states that he cleared the Gig-par-ku of fallen branches. As the excavations progressed we were able to establish that the Gig-par-ku was part of the temple complex dedicated to the moon-goddess, and that it lay close under the southeast end of the ziggurat. Somehow this site of the building had become encumbered with tree branches, and since it was roofed in it likely did not have its own trees -- they perhaps fell from the Ziggurat itself, whose terraces could have had trees planted in soil right above the mud brick, whose waterings would have necessitated the weeper holes.
Lloyd, Setton. 1978. The Archaeology of Mesopotamia: From the Old Stone Age to the Persian Conquest. London: Thames and Hudson Ltd. (Page 153)
Woolley, Leonard. 1954. Excavations at Ur. London: Ernest Benn Limited. (Page 132, when Woolley gets imaginative.)
Elizabeth Carter. ANE 164A Lecture, 8 March 2011. (An at times chilling and humorous discussion of the giparu and the dead entus.)