Innovative wetland management in central Tanzania: case studies

Innovations for Resilience: Small-scale afforestation at wetland border for improved water harvesting in Tanzania

Steven, a farmer in Kalolo District in the central southern highlands of Tanzania, cultivates in the wetland adjacent to a small stream in the centre of a shallow valley characterized by seasonal flows; in the dry season, the channel of flow is less than a meter wide. The wetland and its catchment have been largely converted to agriculture, with some grazing and plantation forestry in the uplands. Steven’s farm is located 1.4 km downstream of the valley head and source area of the stream, which has a stand of original forest (of just 0.35 ha) protected by law.


In 1999 he started preparing terraced fields/ basins in the wetland area close to the stream in order to begin cultivating and to facilitate the irrigation of individual fields. Two years later he invested in a diesel pump to be able to irrigate these fields in addition to some fields just above the wetland area. The pump was expensive to install and he found it costly to use because fuel had to be bought from a distant petrol station. Furthermore, he found that the stream water was often not sufficient in the dry season due to extraction by farmers upstream.

Between 2000 and 2001, Steven started considering the origin of the water that flowed in the wetland stream. He spent time travelling up the catchment to investigate the source. Through his observations of the landscape, he gained an understanding of where water could be found in the landscape, how it flowed, and which natural conditions dominate where water is in abundance. He recognized that water often emerged from under rock outcrops (as springs), but was also abundant at places with natural vegetation (i.e. unconverted wetland areas) like that found at the protected wetland source.

Experimentation and innovation

He began to consider whether he could possibly create those same kind of water-generating conditions on his own farm. He knew he wouldn’t be able to transport to his farm the large rocks associated with springs, but that he could, possibly, establish the same kind of “water-friendly” vegetation on part of his farm. He began to transplant tree saplings found at these naturally wet areas to a piece of land just upslope of his wetland farm that could be described as barren, degraded pasture. He started with Ficus cymocurus but gradually incorporated a mix of trees and shrubs that he had found at those untouched parts of the wetland. The tree he planted most frequently was an endemic wetland species known locally as “mivenge” (Syzygium cordatum). During 2000 and 2001, Steven reforested an area of about 0.3 ha. The forest block extended for 93 m along the upper edge of his wetland farm and for up to 50 m upslope.

Figure 1: Steven in front of the forest block, with Syzygium cordatum in the background


After 4 to 5 years (during 2005 and 2006), Steven noticed that the water table just below the forest had risen. By then the trees were 1.2-1.5 m high. He started planting Grevillea robusta to increase the mulch in the forest. At the borders of the forest he planted sisal to keep livestock out of the forest and to mark the boundaries.  Wetland indicator plants, like ferns, started to grow by themselves.

In 2007 he dug a 0.6 m deep hole at the downslope edge of the forest, adjoining the wetland, and that filled with water after 2 days. The same year, in July (middle of the dry season), he decided to dig a 1.2 m deep storage basin (or collection pool) at the centre of the lower edge of the forest block (at the wetland edge) in order to be able to easily utilize the water. After two weeks, the basin filled to a depth of 0.6 m with water. He also discovered that as the water table was coming up in the forest, some of the original species he planted were no longer coping well (e.g. Syzygium guineese and Pinus patula). In fact, nowadays these species can only be seen close to the top border of the forest (the upper slope) due to the saturated conditions that have since developed along the bottom edge of the forest.

Steven noticed that water was seeping out of the basin he had dug and decided to try to reduce the pressure of the water by creating more basins. Between 2008 and 2010, he built three additional basins that extended along the entire bottom edge of the forest and connected them with plastic pipes. In 2012 he replaced the earthen dam at the lower edge of the basins with an earth-covered brick wall – again in order to reduce seepage losses. He uses the water that fills the basins, which hold a maximum of 120 cubic meters, to flood irrigate 3.6 ha of adjacent wetland fields in the dry season. In the months of May, June, July and August it takes 4-5 days for the basins to refill after they are emptied completely. As the dry season goes on, the recharge rate decreases so that in September, October and November recharge is slowest, and it takes approximately 10 days for the basins to refill after being drained.

Figure 2: A view across the four constructed basins, with the lower edge of the forest block seen on the left (photo taken from point marked with   “  ” in Figure 3)

Figure 3: Plan view of the wetland and Steven’s farm


Towards more sustainable farming?

Today, Steven’s forest is mature, with a dense canopy and a thick layer of mulch on the forest floor. Aside from two beehives kept within it, the forest is unvisited by humans and animals and appears as a natural forest. For Steven, the critical benefit of his planted forest is the precious water it provides him in the dry season when others struggle to source irrigation water and the stream is barely flowing. By regenerating these 0.35 ha of degraded upland with forest, he secures the ability to irrigate 3.6 ha of rotationally fallowed valley bottomland where maize, beans, and – most profitably – tomato are his main crops.

Although not totally conclusive, it is apparent that the forest acts here as a “water trap” in which runoff is minimized and infiltration is maximized. Transpiration of native wetland species is potentially low relative to alternative species, but more importantly the thick canopy cover and ground cover mean that soil evaporation losses are minimized and infiltration is maximised. Without this critical shade, the high water table would incur significant losses of water since it is so close to the soil surface. Nearly all the water entering the forest block is “captured” and is stored in the soil below.

It is the soil hydrology – very saturated conditions – that determines the natural functioning of wetlands. Yet, from Steven’s example we are reminded that soil hydrology is also dependent on the vegetation; and moreover, that even small-scale afforestation using native species can be key to restoring wetlands to a state closer to their natural (pre-utilization) state of functioning. Although results at other sites could depend highly on topography and sub-surface conditions like soil type and sub-surface geology, this innovation could be applicable to many wetland periphery sites, especially on borders of wetlands that have incurred significant losses of natural vegetation and a subsiding water table over the long term. Observations at similar sites will be needed to better understand the phenomenon.


Recorded October 2016

Reporters: Ann-Kathrin Lichtner1, James Ellison1, and Moses Logani2

1 Wetland Action with funding from AKB Stiftung

2 Society for Service to Rural Development in Tanzania (SSRD, a local Iringa-based NGO)