Remote Sensing as a Precision Farming Tool in the Nile Valley, Egypt

Abstract

Detecting stress in plants resulting from different stressors including nitrogen deficiency, salinity, moisture, contamination and diseases, is crucial in crop production. In the Nile Valley, crop production is hindered perhaps more fundamentally by issues of water supply and salinity. Predicting stress in crops by conventional methods is tedious, laborious and costly and is perhaps unreliable in providing a spatial context of stress patterns. Accurate and quick monitoring techniques for crop status to detect stress in crops at early growth stages are needed to maximize crop productivity. In this context, remotely sensed data may provide a useful tool in precision farming. This research aims to evaluate the role of in situ hyperspectral and high spatial resolution satellite remote sensing data to detect stress in wheat and maize crops and assess whether moisture induced stress can be distinguished from salinity induced stress spectrally. A series of five greenhouse based experiments on wheat and maize were undertaken subjecting both crops to a range of salinity and moisture stress levels. Spectroradiometry measurements were collected at different growth stages of each crop to assess the relationship between crop biophysical and biochemical properties and reflectance measurements from plant canopies. Additionally, high spatial resolution satellite images including two QuickBird, one ASTER and two SPOT HRV were acquired in south-west Alexandria, Egypt to assess the potential of high spectral and spatial resolution satellite imagery to detect stress in wheat and maize at local and regional scales. Two field work visits were conducted in Egypt to collect ground reference data and coupled with Hyperion imagery acquisition, during winter and summer seasons of 2007 in March (8-30: wheat) and July (12-17: maize). Despite efforts, Hyperion imagery was not acquired due to factors out with the control of this research. Strong significant correlations between crop properties and different vegetation indices derived from both ground based and satellite platforms were observed. RDVI showed a sensitive index to different wheat properties (r > 0.90 with different biophysical properties). In maize, GNDVIbr and Cgreen had strong significant correlations with maize biophysical properties (r > 0.80). PCA showed the possibility to distinguish between moisture and salinity induced stress at the grain filling stages. The results further showed that a combined approach of high (2-5 m) and moderate (15-20) spatial resolution satellite imagery can provide a better mechanistic interpretation of the distribution and sources of stress, despite the typical small size of fields (20-50 m scale). QuickBird imagery successfully detects stress within field and local scales, whereas SPOT HRV imagery is useful in detecting stress at a regional scale, and therefore, can be a robust tool in identifying issues of crop management at a regional scale. Due to the limited spectral capabilities of high spatial resolution images, distinguishing different sources of stress is not directly possible, and therefore, hyperspectral satellite imagery (e.g. Hyperion or HyspIRI) is required to distinguish between moisture and salinity induced stress. It is evident from the results that remotely sensed data acquired by both in situ hyperspectral and high spatial resolution satellite remote sensing can be used as a useful tool in precision farming in the Nile Valley, Egypt. A combined approach of using reliable high spatial and spectral satellite remote sensing data could provide better insight about stress at local and regional scales. Using this technique as a precision farming and management tool will lead to improved crop productivity by limiting stress and consequently provide a valuable tool in combating issues of food supply at a time of rapid population growth.