Research |
|||
|
water use , cultivation practices , irrigation systems , cultivars , soils , etc . Note : For extracting crop-specific information such as in Figures 2 and 3 , maps showing the distribution of crop types are required . Some of these maps were generated as part of the project , while others were made available by different interest groups .
Both the annual and seasonal crop water use information can assist managers from different water management areas in their planning , since it provides the estimated water use range per crop , per month . Also , given the fact that many new cultivars and crops have been introduced to South Africa over the last 20 years and that little knowledge exits in terms of their water use and crop water requirements , the results from this research contain a wealth of new information .
Mapping irrigated agriculture
Mapping the irrigated areas of South Africa involved a number
|
of steps , each utilizing several datasets . The process followed is summarised in Figure 4 . In the first step , remote sensing and other spatial data were used to produce a map with as high an accuracy as possible . Available field boundary datasets ( based on 2.5m resolution SPOT5 satellite imagery of 2012 ) was updated using SPOT6 satellite data from 2014 / 15 ( at a 1.5m resolution ). Other imagery for 2014 ( e . g . Landsat 8 at 30m resolution ) was also used to identify agricultural expansion since 2012 , eliminate nonagricultural areas and identify potentially irrigated areas such as centre pivots . The annual water use ( or ET ) map ( described in the previous section ) together with an annual rainfall ( P ) map was used to create a map showing the difference between the water use and rainfall ( i . e . ET – P ) for every 250mx250m area in South Africa . Using all these datasets , and applying the assumption that where water use exceeds the rainfall ( i . e . ET – P > 0 ), irrigation is likely to occur , the first version of the |
irrigated agriculture map was generated . The accuracy of this map was estimated to be about 80 %, from which it was concluded that the ET – P approach worked well for most areas . Exceptions were found for : ( a ) some high rainfall regions of the Limpopo province ; ( b ) small irrigated fields ( excluded due to the low spatial resolution of the ET data ); ( c ) field receiving occasional or small amounts of supplementary irrigation ; and ( d ) fallow fields . The initial map consequently needed refinement , which lead to more research and adjustment in the mapping procedures .
For the refined version ( version 2 ) of the map , South Africa was classified into different climatic regions and additional remote sensing derived datasets were obtained . The additional datasets consisted of various indices calculated from high resolution Landsat 8 imagery for different periods ( winter , summer , entire year ) as well as additional outputs from the ETLook model . A machine learning , CART ( Classification and Regression Tree ) analysis was performed on each of the climatic regions to identify rules and thresholds ( decision trees ) that are optimal for each region . The results were subsequently cross-referenced , verified and manually improved for each region . The latest version of the map , showing actively irrigated agricultural areas for the year 2014 / 15 , is shown in Figure 5 .
What is next ? The project is currently entering its final year , during which feedback is being solicited on : 1 . The accuracy of the irrigated agriculture map ;
|
2 . The water use by irrigated agriculture and ;
3 . Water use by individual crops .
This will allow for the finalisation of the calculations on the 2014 / 15 estimated area under and water use by irrigated agriculture . The final task will involve integrating the irrigated agriculture and water use maps with other datasets to assess the water availability within different water management areas , through a process of Water Accounting . Water Accounting describes a catchment ’ s hydrological processes , manageable and unmanageable water flows and their interaction with land use . This information will be useful in assessing possible expansions in irrigated agriculture .
You can assist
You can contribute to this research in a number of ways . First of all , the map shown in Figure 5 , is available for comment and evaluation . Please visit the web portal below to provide inputs on the accuracy of the map . http :// sungis10 . sun . ac . za / fields _ wrc /
Alternatively , if you would like to obtain the data for a larger area for evaluation purposes , please contact the research team . Secondly , if you have information available on the distribution of specific crop types that you would like to share and which could assist in accurately representing a specific crop type , please contact the project team .
|