Trying again (too many windows open before):

Piles of animal dung reveal the location of an ancient Arabian oasis is a for-profit site that aggregates science articles. So I went looking and found this, with more information, but still not the whole story.

Abstract
Aim
Dryland ecosystems, such as those in southern Arabia, are particularly vulnerable to climate and land-use change. Aridification and human subsistence changes at the end of the Holocene Humid Period at 6–4 ka have been used as an iconic example for evaluating such impacts to resilience of arid systems. Although records of ancient environments can provide critical insight into the biotic and abiotic mechanisms that alter ecosystems, traditional archives, such as lake deposits, are not common in southern Arabia after 6 ka; thus, we must use alternative archives.

Location
Yemen.

Taxon
Plants; Rock Hyrax (Procavia capensis).

Methods
We use fossil pollen and stable isotopes (d15N; d13C) of rock hyrax (P. capensis) middens from Wadi Sana, Yemen, to look at changes in ecosystem structure and function across this key transition from 6 ka to the present. A total of 17 middens were radiocarbon-dated and stable isotopes were measured. Of these, pollen was extracted from hyraceum of 14 middens and identified using a light microscope. Fossil pollen assemblages were then compared to existing modern pollen samples from throughout Arabia.

Results
During the mid-Holocene from at least 6 to 4.7 ka, the pollen flora reflects a landscape with abundant tropical trees. These included foundational woody taxa, such as Terminalia and Boswellia sacra (frankincense), which had a strong positive influence on local hydrology and the economy, respectively. Increased charcoal abundances also suggest that wildfire occurred periodically.

Main conclusions
Connections with archaeological evidence suggest a strong link between human management and the presence of Terminalia woodland during the mid-Holocene. These may have promoted increased groundwater storage and ponding as regional rainfall was decreasing. The mid-Holocene expansion of Boswellia sacra (frankincense) may have helped support resin trade that became a critical export from the region. Fires were more common than today, suggesting semi-arid but continuous vegetation cover across what is now bare ground. Finally, after 1 ka, increased sedge abundances at the expense of grasses and trees suggest the development of desert conditions.

It turns out this may be a very old story, because another link I found is this: https://www.researchgate.net/publication/281474114_Early_Arabian_Pastoralism_at_Manayzah_in_Wadi_Sana_Hadramawt

I've accidentally deleted this page once, so I'll submit this message and sort it out later. What I find interesting is the archaeological information found in the middens of small desert animals. The same is true of packrat middens in the US desert southwest.