*2.2. Description of WEAP Model*

The water evaluation and planning model developed by the Stockholm Environment Institute (SEI) is a decision support system (DSS) used for the integration of water resources management and planning. It is easy to use for water planning and scenario assessment. WEAP simulates water balance for water demand, supply, and storage on a monthly basis and it allows the assessment of water resource management policies between different sectors (agriculture, industry, tourism). It can be applied at a catchment level as well as other more complex levels such as regional and country levels [36].

Within the WEAP model, different agricultural catchment calculation methods can be used. In this study, we used the WEAP-MABIA method version 1.0.1 [35] to simulate crop water requirement, crop yield as well as agricultural management plans under different climate conditions. The selection of this method was based on the fact that it has been applied by scientists, engineers, and resource managers to simulate runoffs, infiltration, and percolation processes resulting from natural rainfall, irrigation scheduling, and crop yield reduction [35,37–39]. The WEAP-MABIA method calculate evapotranspiration using the 'dual' crop coefficient kc method (Kc = Ke + Ks Kcb), as described in Allen et al. [40], whereby the Kc value is divided into a 'basal' crop coefficient, Kcb, and a separate component, with Ke, representing evaporation from a shallow soil surface layer. The basal crop coefficient represents actual ET conditions when the soil surface is dry but sufficient root zone moisture is present to support full transpiration.

The reference evapotranspiration (ET0) for the Olifants catchment was calculated using the modified Penman-Monteith equation recommended by Allen et al. [40]. The

equation utilizes some assumed constant parameters for a clipped grass reference crop. It was assumed that the definition for the reference crop was a hypothetical reference crop with a crop height of 0.12 m, a fixed surface resistance of 70 s m−<sup>1</sup> and an albedo value (i.e., a portion of light reflected by the leaf surface) of 0.23 [41]. The equation used for calculating ET0 is given below:

$$ET\_0 = \frac{0.408\Delta (R\_N - G) + \lambda \frac{900}{T + 273} \cup \left(e\_s - e\_a\right)}{\Delta + \gamma (1 + 0.34 \cup\_2)}$$

where *RN* is the net radiation at the crop surface, *G* is the soil heat flux density, *T* represents the mean daily air temperature at 2 m height, *U2* is the wind speed at 2 m height, *es* is the saturation vapor pressure, *ea* is the actual vapor pressure, (*es–ea*) represents the vapor pressure deficit of the air, Δ is the slope vapor pressure curve, and γ represents the psychometric constant.

The following climate parameters such as daily temperature (minimum and maximum), average relative humidity, 2-m wind speed, and solar radiation were used to estimate the current and projected reference crop evapotranspiration.

The performance of the WEAP-MABIA was verified by calibrating and validating observed crop yield data for Mpumalanga province where the Olifants catchment is situated, as there was no recorded crop data for the catchment. The data was obtained from the Department of Agriculture, Fisheries, and Forestry (DAFF). The efficiency of the model performance was determined by comparing the observed against the simulated crop yield using two verification statistics such as Coefficient of Determination (*R2*) and Nash-Sutcliffe Efficiency (NSE). The values of *R2* ranges between 0–1, values higher than 0.5 are considered acceptable. While NSE ranges between −∞ and 1.0, where NSE = 1 indicates a perfect match of simulated and observed yield. An efficiency of 0 shows that the model prediction is as accurate as the mean of the observed data, while an efficiency less than 0 shows that the observed mean is a better predictor than the model. For more detail on the procedure and statistical equation used for the calibration and validation of the model, readers should consult Olabanji et al. [29]. We calibrated and validated the WEAP-MABIA crop model using observed crop yield data for the period of 1995–2000 for calibration and 2001–2004 for validation. The results presented in Table 3 show that the simulated crop yield perfectly agrees with the observed yield with NSE ranging between 0.97 to 0.99 during calibration and 0.87 to 0.96 during validation. The coefficient of determination (*R2*) ranged between 0.98 to 1.0 for calibration and 0.95 to 0.98 during the validation process. The agreement between the simulated and observed crop yield indicates the capability of the crop model to simulate future crop yields.


**Table 3.** Model calibration and validation result using yearly simulated and observed crop yield.
