Mazowe Catchment is one of the seven water management catchments in Zimbabwe. It lies between 16.470∘ S (latitude) and 18.240∘ S (latitude) and between 30.680∘ E (longitude) and 33.000∘ E (longitude). The catchment has a total area of 38 005 km2 and is in the northern part of the country. There are 30 major dams built along the main river and some of the tributaries. There are 13 functional gauging stations within the catchment. The catchment is composed of twenty-four (24) hydrological sub-zones. Rainfall for the Catchment averages 500 mm to 1200 mm yr-1 while the mean annual runoff is 131 mm with a coefficient of variation (CV) of 126 % (Zinwa, 2009). There are 17 water quality monitoring sampling points within the catchment. Six (6) sampling water points were selected for the study, these points were selected basing on the continuity of the available data meaning to say the points had less missing gaps.

Chimhanda Dam, in Lower Mazowe Sub-Catchment, was chosen for the assessment of the levels of sedimentation in resevoirs. Chimhanda Dam is located on the confluence of Runwa and Mwera Rivers, which are tributaries of Mazowe River. The dam is located in hydrological sub zone, DM1 with latitude 16∘40′, and longitude 32∘06′. The dam was completed in 1988 with a design capacity of 5.2 × 106 m3 and covering a catchment area of 68.7 km2, which mainly consists of communal lands.

Remote sensing images of the study area were downloaded from Landsat TM 4-5, LandSat LE7 and LandSat 8 scene from the website (https://glovis.usgs.gov/) for paths 168–170 and rows 71–73 for the period, 2000, 2005, 2008 and 2014. The study period was selected basing on the major changes in land use and land cover as a result of the Zimbabwe land reform programme which started in year 2000 and peaking during year 2003. The composite map was obtained by gluing and merging the tiles from the different scenes which was performed using Integrated Land and Water Information System (ILWIS) software. Normalized Differences Vegetation Index (NDVI) was extracted from the images and this was used to calculate the cover and management practice factor (C factor). SRTM DEM of 90m resolution was obtained from Earth Explorer website (http://EarthExplorer.com). The DEM was used to calculate the slope length and slope steepness factor (LS factor). The erodibility factor was obtained from the soil map of Zimbabwe that was downloaded from website (www.fao.org/geonetwork. Rainfall data was obtained from the Zimbabwe Meteorological Services Department and was used to come up with the erosivity factor. The grab sample method was used to obtain samples for sediment loads from the flow gauges for the 2014/2015 rainfall season. Historical data for sediment loads from flow gauging stations were obtained from the Zimbabwe National Water Authority (ZINWA). Historical water quality data for selected water quality monitoring stations was obtained from Environmental Management Agency (EMA).

Siltation historical data for selected reservoirs within the catchment was obtained from ZINWA. A bathymetric survey was done to assess the level of sedimentation of Chimhanda dam which lies within the Mazowe Catchment.

The Revised Universal Soil Erosion (RUSLE) model was used in this study to calculate the soil that was lost from the catchment. The model predicted the long term annual loss from the basin and is given by Eq. (1; Renard et al., 1997): A=R×K×LS×C×P where: A= annual soil loss (t ha yr-1); R= rain erosivity factor (MJ mm ha-1 h-1); K= soil erodibility factor (t ha h MJ-1 mm-1); LS = slope length and slope steepness length (m); C= land cover and crop management; P= management practice.

The estimated soil loss was calculated from Eq. (6). A=R×K×LS×C×P where A= Average Soil Loss (t ha yr-1); R= R factor map; K= K factor map; LS = LS factor map; C= C factor map; P= P factor.

Soil loss from the catchment cannot be taken as sediment contribution to a river flow system since it does not account for deposition that occurs along the path (de Vente et al., 2011). Therefore the estimated soil loss was multiplied by the sediment delivery ratio (SDR) to obtain the sediment yield of the catchment. Sediment delivery ratios represent the fraction of the total soil loss that is washed into rivers and was calculated from using Eq. (7; USDA, 1972). SDR=05656CA-0.11 where SDR = Sediment Delivery Ratio; CA = Watershed Area, km2.

After determination of the sediment delivery ratio, the average sediment yield was determined using Eq. (8), by Wischmeier and Smith (1978) SR=SDR×A where: SR = Sediment yield (t ha yr-1); SDR = Sediment delivery ratio; A= Average soil loss (t ha yr-1).

Seven out of the thirteen functional ZINWA gauging stations were selected for collection of sediment concentration samples, (Fig. 2). Water samples were collected from various flow gauging stations for the period November 2014 to March 2015 using grab sampling method. An average of 20 samples was collected from each gauging station. The samples were analyzed in a laboratory using the weighing and filtration method inorder to determine the sediment concentration in  mg L-1. The sediment load was determined by multiplying the sediment load at a particular gauge by the area of influence.

A bathymetric survey was carried out at Chimhanda dam to determine the siltation level of the dam. The survey was conducted using a SonTek River Surveyor system and a theodolite to come up with a basin survey map. The plotted map was digitized to get the surface area between contours. The area was then multiplied by the contour interval to get the volume of each contour using Simpson's formula. Vcontour=A1+(A1×A2)1/2+A23 where Vcontour= contour Volume (m3); A1= Area1 (m2); A2= Area2 (m2).

The calculated volumes for each contour were accumulated to get the new capacity of the dam.

Water quality historical data from six selected water sampling points were obtained from the Environmental Management Agency records. The points were selected basing on the consistency of the data and less gaps. The trend for the selected water quality parameters, TSS, DO, NO3, pH, TDS and turbidity for the corresponding predicted sediment yield years were analysed. The water quality parameters were compared with the Environmental Management Agency, Effluent and Solid Waste Disposal Regulations Statutory Instrument number 6 of 2007 (EMA, 2007) shown on Table 1. Water quality data for Chimhanda water supply treatment plant from the Zimbabwe National Water Authority was also related to the change in level of sedimentation of Chimhanda Dam.

Selected water quality parameters were correlated with the corresponding sediment yield to determine the relationship between sediment yield and water quality. Pearson Correlation was used to estimate the strength of relationship between sediment yield and some physical water quality.

The erosivity map (R factor) depicts rainfall energy in the various areas within the catchment. The rainfall erosivity ranges between 200 and 500 (MJ mm ha-1 h-1 yr). A greater part of the catchment is averaging a rainfall erosivity value of 276 MJ mm ha-1 h-1 yr. Highest rainfall erosivity values are in the Kairezi sub-catchment of the study area. Erodability (K factor) values range from 0 to 0.5 t ha-1yr-1 MJ-1 mm-1. Some parts in the south west of the catchment have high values of erodibility, the highest K value is dominated by very fine sand with silt particles which give rise to higher soil erodibility (Kamaludin et al., 2013). Highest slope length steepness (LS) values of 16 m were found in the eastern part of the catchment. The northern part of the catchment experiences average LS values of 7 m.

Cover Management and Practice (C factor) ranges from -1 to 1. The C factor is depicted by NDVI which is a function of photosynthesis. The C factor is high in the eastern highlands area, some parts of Kairezi sub-catchment and along the Mufurudzi area. This is possible because the Kairezi area experiences high rainfall while Mufurudzi being a game park, has more vegetation. High C values were observed with high vegetation cover (Pettorelli et al., 2005). The P value of 1 was observed in Lower Mazowe and Upper Rwenya subcatchment areas respectively. Areas like Mufurudzi and parts of Kairezi experienced an average P value of 0.8.

The average annual soil loss for the different years are shown on Fig. 3 and the temporal variation of actual soil loss is tabulated on Table 2.

The estimated soil loss from the catchment ranges from 0 to 203 t ha yr-1. The estimated average soil loss is 54 t ha yr-1. The highest soil loss is being experienced in Middle Mazowe and Nyagui sub-catchments. High soil loss in Middle Mazowe sub-catchment can be associated with high gold panning activities in the Mazowe valley whilst in Nyagui sub-catchment, this can be associated with high crop farming activities in the area. When the soil is made loose, its structure is altered hence increase in erodibility.

The average soil loss value of 54 t ha yr-1 concurred with other studies that were carried out in the country. Whitlow (1986) found that 76 t of soil is lost per hectare per year through soil erosion in most parts of the country. Another study by Mutowo and Chikodzi (2013) “Erosion hazards mapping in the Runde Catchment”, concluded that most of the areas in the catchment fall in the category range of 0–50 t ha yr-1. Makwara and Gamira (2012) reported that the most serious type of erosion being sheet erosion, is estimated to remove an average 50 t ha yr-1 from Zimbabwe's communal lands.

There is no uniform trend in the soil loss over the study period. In year 2000, the soil loss was 54t ha yr-1. The increase of soil loss in 2005 from the 2000 figure can be explained by the land reform programme which started in year 2000 and was at its peak around 2003. There was a lot of deforestation as new farmers were clearing land for agricultural purposes (Mambo and Archer, 2007). Some areas which were meant for animal rearing and forest were also converted to crop agricultural areas. In 2008 the soil loss rate decreased to 36 t ha yr-1. The reduction can be attributed to a number of factors. Year 2008 experienced low erosivity as a result of low rainfall, which could also lead to low annual soil loss. The economy of Zimbabwe was almost at a stand-still during the period 2007 to 2009, with an inflation rate record of over 231 million percent in July 2008 (Hanke and Kwok, 2009). As a result there was less farming activities during that period.

The soil loss rate almost doubled in year 2014 as compared to 2008. This can be explained by the increase in the number of small scale gold miners in the catchment and alluvial gold panning in streams and rivers. Mining within the catchment is not only limited to the river beds and banks, miners are also targeting the inland areas of the catchment such as the Mazowe Valley and Mufurudzi Game Park and as a result, soil erodibility is also increasing. Increase in rate of unemployment caused by low capital investment and continuing closure of industries in Zimbabwe has resulted in the population resorting to other sources of income such as illegal gold mining and alluvial gold panning. A study that was carried out in the Lower Manyame sub – catchment along Dande River, on the analysis of the implications of cross- sectional coordination of the management of gold panning activities and its impacts (Zwane et al., 2006), identified river bed gold panning activities as a cause of degradation of river channels and banks as well as accelerated erosion and siltation in many areas of Zimbabwe.

Table 4 summarizes the results of the silt survey for Chimhanda Dam.

The capacity of the dam decreased from the original 5.200 × 106 to 3.470 × 106 and 3.194 × 106 m3 in years 2003 and 2015 respectively. There was 33.3 % loss in storage up to 2003 and a storage loss of 38.6 % up to 2015 in reference to the design capacity. Figure 5 shows reservoir capacity changes over the years.

Results from the 10 samples of river inflows collected gave an average sediment concentration of 39 mg L-1. There was a decrease in trend in the sediment concentration as the rainfall season was progressing. The sediment accumulation in the reservoir was found to be 330 t yr-1 and the specific sediment yield was 226 t km-2 yr-1. The usefulness of Chimhanda Dam has reduced from the initial design life of 50 to 37 years. A study by Godwin et al. (2011) on Chesa Causeway Weir in the same catchment gave a specific yield of 510 t km2 yr-1. The usefulness of the dam was reduced from the initial design life of 50 to 25 years.

Generally the results support the existing literature that most of the small to medium dams are being affected by high rates of sedimentation in Zimbabwe (Sawunyama et al., 2006; Godwin et al., 2011; Dalu et al., 2013; Chitata et al., 2014).

Figures 6 to 10 show the results for the historically measured water parameters for the years 2000, 2008, 2010 and 2014.

Pearson correlation was used to determine the relationship between two water quality parameters and sediment yield. The parameters are related to sediment yield and these are turbidity and suspended solids, (Bhadja and Vaghela, 2013). The results are shown on Table 5.

There is a close positive relationship between turbidity and sediment yield of 0.63. Total Suspended Solids have a positive relation of 0.64. The results show that the two water quality parameters can be indicators of sedimentation. Higher values of total suspended solids and turbidity indicate higher values of sediment yield in water bodies.

The variation in two selected reservoir water quality indicators, turbidity and pH was determined. A significant variation in turbidity was noted for the years 1988 to 2014 with a p-value of 0.02, which is less than 0.05. The mean value of turbidity for 1988, 2003 and 2014 are 15.6, 34.4 and 41.7 NTU respectively. The values show an increasing trend in turbidity which may be as a result of anthropogenic activities occurring upstream of the dam which has resulted in increased soil erosion hence more sediment on the reservoir. pH results showed that there is no significant change between 1988 and 2014 (p-value > 0.05). This might be because pH is a chemical parameter that is determined by the type of sediments that are carried into the river not the sediment load.

The objective of the study was to analyze sedimentation and its impact on river system and reservoir water quality in Mazowe Catchment, Zimbabwe. The study's major findings are,

The catchment is generally experiencing moderate soil loss, however there are some pockets under high and extreme soil loss which gave rise to siltation and water quality deterioration of available water bodies.

There is reduction in capacity of reservoirs within the catchment due to sedimentation. Chimhanda Dam capacity has been reduced by 38.6 % from the initial design capacity as of April 2015, reducing the usefulness of the dam to 37 years from the initial design life of 50 years due to siltation.

Sediment deposition in rivers affects water quality and water quality parameters can be used as an indicator for sedimentation.

In order to curb erosion and sedimentation in rivers and reservoirs, there is need to develop and implement an integrated water resources management plan by all stakeholders.

In order to maintain the lifespan of Chimhanda dam, best land management practices such as contour ploughing and re-forestation should be employed around the catchment area of the Dam and this is an area that could be considered for further research.