Important to consider rainfall, soils, topography and native or desirable (non-invasive) species options. I was present some decades ago, when Moshe Cohen (Israel) presented work to establish native cedars in ephemeral watershed that was more or less desert. In reforestation of a former forest that had long been deforested and basically functioned as a desert, the results of establishing the forest, dry streams eventually converted back to their original perennial nature. They established individual tree planting depressions across the landscape to capture rainfall and funnel water to tree. Each planted depression is also ~2-3 feet closer to water table. Selecting the best and dry hardiness species, selecting best time to plant in a moister period, avoiding periods of extensive drought, augmenting water if needed until trees get established rather than letting them die. A soil scientist could also help evaluate if there are nutrient or other soil limitations. Mulch and fertilization may be needed. You might also try PAM, a granular treatment to be added with each tree, that when rain hits, water is absorbed into a clear goo that develops, and helps slowly release water to nearby plants. Also, some tree species do best with mychoriza. Dry deforested soils may have lost their presence if once available. Native grass or other species may have also been lost if once present. There are a number of technical contacts relative to botany, soils and forestry that could help offer advice or make wise selections. If there are options to use drip or other conservative irrigation, that can be considered, and would be helpful if developing tree or plant nursery, but recognize that too much watering may prevent development of deep roots needed to sustain plant. In modifying the surface to capture and retain more water, one must be sure to avoid actions contributing to overland flow and excessive erosion. Some of our actions in treating gullies in SC, USA (various researchgate papers) might be useful In understanding recovery may take several actions and periods of maintenance to achieve success.

It is going to be more difficult to accomplish if selecting an especially dry year or period, or during a drought. If you have periods or years of forecast with above normal rain and improved frequency, those are best. Sometimes it may be helpful to also have other plant cover, so identifying grass or other species that will be compatible with trees that do not do not use too much water and add organics and nutrients helpful to trees (such as nitrogen fixation). If developing a nursery, it may be helpful to select seed from the desired species which have been able to survive and more or less thrive under harsh conditions (exhibited drought tolerance).

The above answer of Mr Hansen is very good, I would just add that this method is also quite well efficient to optimize the infiltration of water in the local aquifer , while it may preserve the soil from the erosion and limit the runoff factor in the benefit of the trees alimentation, as well as to attenuate the possible flooding impact...

Water harvesting to concentrate it to the area of tree establishment very necessary as well as initial favorable planting period and some inititial irrigation. In addition the mycorrhizal condition of the tree species can be critical when trees are conclusively mycorrhizal they are much more able to efficiently use scarce water resources. A fully mycorrhizal plant can effectively extend the area for drawing water up to thousands of times the nonmycorrhizal condition.

I think the above answers are quite on the mark. I would only add that experience would indicate that understanding the soil conditions is critical. Reforestation and water harvesting are complex. Depending on the sediment matrix and amount of water employed, temperatures and air movement could lead to excessive salts being leached out of the soils and result in weakened growth of a replacement forest. Also, the chemistry of the extracted water should be understood very clearly.

I appreciate the explanation of Dr. W F Hansen. Infiltration and water retention depends on the hydraulic properties (porosity & permeability) of the soils. Terrain slope, vegetative cover; sun radiation; atmospheric temperature; etc., also plays an important role in the effective conservation of harvested water. Basically water harvesting is carried out for multiple purposes like improving soil moisture; ground water recharge; revitalization of rivers; storage reservoirs / ponds for irrigation; etc. There are many structures & methodologies available for water harvesting like pits; trenches (continuous & discontinuous); mulching of soil; contour bunds; check dams; recharge ponds; etc.

The desert soils can be sandy and bereft of organic matter. Beside the water harvesting techniques the import of organic amendment will be useful along with mycorrhizal provision. A compost clay mixture could be the source of much greater water retention and use in the planting hole along with other techniques already expounding on. In our areas peat moss has been used with benefit and is helpful in neutralizing an alkaline soil condition which can be detrimental for many trees.

The construction of keylines will modify the topography to facilitate the retention of water from the land. Keylines become therefore the restoration starting points. Knowledge of native tree, shrub and grass/forbs species is essential as the successful establishment of a native plants community will slowly reclaim a desert soil through an improvement of its structure and fertility to accommodate (eventually) agronomic species too.

My research group and I wrote a paper on this for house-scale water sheds, considering the field capacity and the wilting point of the soils.

The paper does not go much to discuss the nutrient loading nor the soil conditioning, but I hope this resource helps.

Thank you all for answers that I think are very interesting.

Thanks to Osvaldo Broesicke for his paper and to Arash Malekian for his accurate opinion.

I think Bruno Borsari is quite right when he talks about the keylines as starting points of the restoration; we call this "oasification" (a strategy to reverse the process of desertification due to soil aridity by modifying the relief of the hillsides).

I also find interesting the note of Paul Reed Hepperly, about the need to improve desert soils when they are sandy and poor in organic matter, although not all arid soils are sandy, of course.

Thanks also to William F. Hansen for his extensive and complete response, which talks about most of the factors to be taken into account. I know the forest restoration works carried out in Israel so successfully. W.F. Hansen is a person with a lot of experience in the subject and I will read his papers about it.

The contribution of Hamed Ben Dhia is also very useful to complete what was said by W.F. Hansen. As well as the contributions of Paul Reed Hpperly, Henry Hidell and N B Narasimha Prasad.

Now I raise two new related questions:

-Do you think that it is important to quantify well the "water harvest", for the correct design of a water collection system for forest restoration?

-Is it also very important to know precisely how much water is retained in the soil and how much is lost by percolation?

Thank you so much.