Modelling Evapotranspiration of Cannabis sativa L. using a UAV in the KwaZulu-Natal Midlands, South Africa.
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
The recognition of Cannabis sativa L., as a high-value crop, combined with anecdotal evidence of its successful cultivation, has led to suggestions of its significant potential for small-scale emerging farmers in rural areas of KwaZulu-Natal. However, to ensure the feasibility and sustainability of this activity, it is necessary to investigate its impact on water resources in areas that are already water scarce. The South African National Water Act (No. 36 of 1998) mandates the regulation of land-based activities that reduce streamflow by declaring the crop as a streamflow reduction activity (SFRAs). While it is widely known that hemp is a water-intensive crop, no field-based measurements of its evaporation (ET) exist in South Africa. To extrapolate field measurements from a point measurement to assess spatial variation in water use, a remote sensing modelling approach was applied to derive ET using multispectral drone imagery. QWaterModel analysed thermal images acquired from an Unmanned Aerial Vehicle (UAV) over a single growing period of hemp in KwaZulu-Natal, South Africa. The QWaterModel estimates of ET (ET QW ) were compared to eddy covariance ET (ET EC ). A total ET QW of 24.2 mm was modelled, while the EC system measured 16.9 mm over the same five days over the growing season. ET QW variation corresponded well with Normalized Difference Vegetation Index (NDVI), however ET QW estimates at the beginning of the season, and after harvest, were more representative of the surrounding soil surfaces and grass cover than the hemp plants. A strong correlation was observed between QWaterModel and ground-based EC measurements. The lack of canopy closure affected the estimation of ET, as the single-source QWaterModel is unable to differentiate heterogeneous canopies. These results provide the first water use and crop coefficient estimates of hemp in South Africa and data required to assess the streamflow reduction activity of hemp.