DECLARATION 2: PUBLICATIONS
4. PERFORMANCE ASSESSMENT OF THE IMPROVED CONTINUOUS
4.4 Catchments Used for Verification
4.4.1 General climatic and physiographical characteristics
The Mean Annual Precipitation (MAP) values for the catchments were obtained using the national rainfall database and Geographic Information System (GIS) grids, developed by Lynch (2003). All GIS analyses were performed using the ArcGIS 10.4 software (ESRI, 2016).
Catchment areas were obtained from publications for the research catchments (Smithers and Schulze, 1994a; Smithers and Schulze, 1994b; Scott et al., 2000; Gush et al., 2002; Royappen, 2002; Royappen et al., 2002; Lorentz and van Zyl, 2003), and from the DWS website for gauging weirs monitored by DWS. Verification of all catchment areas was performed via the following steps: (i) Google Earth was used to identify and confirm the exact location of the streamflow gauging weirs, (ii) ArcGIS 10.4 was used to delineate the catchments, i.e. using the co-ordinates of the verified gauge locations and 1:50 000 topographical map sheets, available from the CWRR national GIS database obtained from the Chief Directorate of National Geo- spatial Information (CDNGI, 2013), formerly the Chief Directorate of Surveys and Mapping
Calculate Geometry function in ArcGIS 10.4, (iv) these calculated areas were compared to those obtained from the sources listed above, and (v) corrections made if required, i.e. to the areas provided from the aforementioned sources. Similar to the MAP, the mean catchment altitude and slope was calculated in ArcGIS 10.4 using a 20 m Digital Elevation Model (DEM) available from the CWRR national database, also sourced from the CDNGI (2013). In addition to MAP, catchment areas, and mean catchment altitude, specific information about land cover and soils, as summarised in Table 4.1, was obtained from the literature (Smithers and Schulze, 1994a; Smithers and Schulze, 1994b; Scott et al., 2000; Gush et al., 2002; Royappen, 2002;
Royappen et al., 2002; Lorentz and van Zyl, 2003). This is the most accurate and detailed information available for the research catchments at the time of data collection and was used in preference to the default land cover (ARC and CSIR, 2005; DEA and GTI, 2015) and soils maps (Schulze, 2012), suggested for use with the CSM system developed (Chapter 3). DWS Gauges X2H026 and X2H027 fall within the Mokobulaan research catchment area and catchment X2H026 was one of the catchments used by Royappen (2002) and Royappen et al.
(2002) for improved parameter estimation in streamflow predictions using the ACRU model.
For the remaining two DWS gauges (A9H006 and V1H032) detailed information was not available and therefore the default land cover information from the NLC 2000 map was used, and the SCS-SA soils groups were obtained from the national SCS-SA soil group map developed by Schulze (2012), i.e. the recommended default information to use with the CSM system developed. In addition, in the absence of detailed land cover information for X2H027, the NLC 2000 information was also used for this catchment. However, the soils information was the same as that obtained for X2H026, i.e. from the literature describing the Mokobulaan area.
As alluded to in Section 4.2, Weddepohl (1988) delineated South Africa into four rainfall intensity distribution regions, and developed synthetic distributions of daily rainfall for each region. Region 1, with a Type 1 rainfall distribution, has the lowest rainfall intensity with rainfall more uniformly distributed throughout the day, while Region 4, with a Type 4 rainfall distribution, has the highest rainfall intensity with the majority of the daily rainfall falling within an hour (Weddepohl, 1988; Schulze, 1995). Using a map of the rainfall intensity distribution regions for South Africa obtained from Schulze et al. (2004), the rainfall intensity region for each catchment was identified. This was required to calculate Ī30, i.e. the 2-year return period 30-minute rainfall intensity, as needed to estimate the lag time using the Schmidt
and Schulze (1984) lag equation (Equation 4.4). In order to calculate Ī30 an estimate of the 2- year return period maximum 1-day rainfall is multiplied by a multiplication factor defined for each region, available from Schulze (1995). A map of expected 1-day maximum rainfall values for South Africa, i.e. for the 2-year return period, is also available from Schulze (1995). This map can be used with the multiplication factors to calculate Ī30, however, the mapped values are very generalised. Consequently, the 2-year return period maximum 1-day rainfall was calculated from the daily rainfall record used to model each catchment, by extracting the AMS, fitting the GEV distribution with L-moments (Hosking and Wallis, 1997), and extracting the 2-year return period maximum 1-day rainfall from the distribution. The lag time was then calculated using Equation 4.4.
All the information summarised in Table 4.1 is required to apply the CSM system developed.
Much of the information in Table 4.1 is also required to apply the current default implementation of the ACRU model, however, excluding the revised SCS-SA land cover class and the SCS-SA soil group. The ACRU land cover class assigned to the revised SCS-SA land cover class, i.e. for application in the CSM system developed, is also applicable to the current default implementation of the ACRU model. In the current default implementation of the ACRU model, however, the QFRESP (Table 4.1) and SMDDEP (Table 4.1) parameters are not determined based on the SCS-SA CNs, as performed for the CSM system developed. Instead, these parameters are set to default values. QFRESP is set at 0.3 for all catchments and SMDDEP is default to the depth of the topsoil. In addition, in the current default implementation of the ACRU model, the soil parameters required as input to the ACRU model are obtained from a national soils map developed by Schulze and Horan (2008). The soil parameter values obtained from this map for each of the verification catchments, required to apply the current default implementation of the ACRU model, are provided in Table 4.2. When applying the CSM system developed, these soil parameters are obtained from Table 3.4 based on the SCS-SA soil group as defined in the CSM system developed.
Table 4.1 Climatic and physiographical characteristics of the selected verification catchments required to apply the CSM system developed
Catchment Area
(km2) MAP (mm)
Mean Altitude
(m)
Revised SCS- SA Land Cover Class
Treatment
(Class) Type Hydrological Condition
ACRU Land Cover Class Assigned to Revised SCS- SA Class (COMPOVEG
Number / Source)
Mean Slope (%)
SCS-SA Soil
Group CN
QFRESP SMDDEP
Rainfall Intensity
Region
(mm/h) Ī30
Schmidt- Schulze Lag (h) Cedara
(U2H020) 0.26 1093 1106 Unimproved
(Natural) Grassland
2 = in fair
condition Fair UNIMPROVED
GRASSLAND (5060103) 11.00 A/B 61 0.59 0.25 3 49.52 0.54
DeHoek / Ntabamhlope
(V7H003) 0.52 870 1497 Unimproved
(Natural) Grassland
2 = in fair
condition Fair UNIMPROVED
GRASSLAND (5060103) 14.60 B/C 75 0.91 0.25 3 51.49 0.47
Jonkershoek - Lambrechtsbos
B (G2H010) 0.73 1074 517 Forests &
Plantations Humus depth
> 100mm loose or friable /
Site prep pitting FOREST PLANTATIONS
GENERAL (Schulze, 2013) 36.39 A/B 33 0.3 0.45 2 39.54 0.64
Cathedral Peak
IV (V1H005) 0.98 1264 2011 Unimproved (Natural) Grassland
3 = in good
condition Good UNIMPROVED
GRASSLAND (5060103) 32.70 A/B 51 0.37 0.25 4 81.89 0.47
DeHoek / Ntabamhlope
(V1H015) 1.04 943 1512 Unimproved
(Natural) Grassland
3 = in good
condition Good UNIMPROVED
GRASSLAND (5060103) 17.00 B 61 0.59 0.25 3 56.58 0.58
Cedara
(U2H018) 1.31 946 1269 Forests &
Plantations Humus depth
> 100mm loose or friable /
Site prep pitting FOREST PLANTATIONS
GENERAL (Schulze, 2013) 23.30 B 47 0.3 0.26 3 47.70 0.67
Zululand
(W1H016) 3.30 1121 260 Unimproved
(Natural) Grassland
3 = in good
condition Good UNIMPROVED
GRASSLAND (5060103) 13.20 B 61 0.59 0.25 1 34.68 1.74
X2H026 13.82 978 1450
Forests &
Plantations (24%)
Humus depth
50 - 100mm Fair/Intermediate
site prep FOREST PLANTATIONS GENERAL (Schulze, 2013)
30.78 A/B 51 0.37 0.25 3 64.68 1.11
Unimproved (Natural) Grassland (76%)
3 = in good
condition Good UNIMPROVED
GRASSLAND (5060103)
A9H006 16.00 1708 1055 Forests &
Plantations Humus depth
> 100mm loose or friable /
Site prep pitting FOREST/NATURAL
FOREST (5020101) 32.34 B/C 52 0.39 0.25 2 63.76 2.16
V1H032 67.80 982 1571 Unimproved
(Natural) Grassland
3 = in good
condition Good UNIMPROVED
GRASSLAND (5060103) 26.50 C 74 0.89 0.25 3 79.55 1.71
X2H027 77.16 1026 1546
Forests &
Plantations (87%)
Humus depth
50 - 100mm Fair/Intermediate
site prep FOREST PLANTATIONS GENERAL (Schulze, 2013)
30.10 A/B 51 0.37 0.25 3 64.68 2.16
Unimproved (Natural)
Grassland 3 = in good
condition Good UNIMPROVED
GRASSLAND (5060103)
Table 4.2 ACRU soils input information obtained for each verification catchment from the national soils map developed by Schulze and Horan (2008)
Catchments (kmArea 2) DEPAHO (m) (m.mWP1 -1) (m.mFC1 -1) (m.mPO1 -1) ABRESP/BFRESP DEPBHO (m) (m.mWP2 -1) (m.mFC2 -1) (m.mPO2 -1) Cedara (U2H020) 0.26 0.30 0.172 0.275 0.406 0.37 0.67 0.217 0.333 0.427 DeHoek /
Ntabamhlope
(V7H003) 0.52 0.30 0.150 0.240 0.422 0.42 0.63 0.208 0.292 0.413 Jonkershoek -
Lambrechtsbos B
(G2H010) 0.73 0.26 0.115 0.201 0.445 0.46 0.12 0.121 0.211 0.443 Cathedral Peak IV
(V1H005) 0.98 0.30 0.134 0.224 0.439 0.39 0.55 0.156 0.248 0.410 DeHoek /
Ntabamhlope
(V1H015) 1.04 0.30 0.137 0.223 0.433 0.44 0.72 0.197 0.268 0.406 Cedara (U2H018) 1.31 0.30 0.170 0.270 0.410 0.37 0.62 0.212 0.324 0.425 Zululand
(W1H016) 3.30 0.30 0.120 0.212 0.462 0.36 0.12 0.106 0.205 0.439 X2H026 13.82 0.30 0.173 0.280 0.399 0.38 0.67 0.199 0.317 0.425 A9H006 16.00 0.30 0.169 0.277 0.404 0.38 0.85 0.212 0.338 0.431 V1H032 67.80 0.30 0.144 0.233 0.432 0.38 0.36 0.177 0.265 0.416
X2H027 77.16 0.30 0.174 0.282 0.398 0.37 0.60 0.196 0.314 0.425