No abstract

No abstract

Spatially distributed air temperature (Ta) data are essential for environmental studies. Ta data are collected from meteorological stations of sparse distribution. This problem can be overcome by using remotely sensed datasets at different scales. This study used land-based temperature measurements and satellite data for estimating Ta distribution over the United Arab Emirates. Landbased Ta data from 11 weather stations during 2003 to 2019 were used with MODIS Aqua LST for both daytime (LSTd) and nighttime (LSTn) data. The results indicate a significant correlation between LST and Ta with regression coefficients R 2 > 0.94/0.96 and Root Mean Square Error about 1.75/0.97 C of LSTd/T max and LSTn/T min , respectively. Large variability was observed between the daytime and nighttime mean temperature distribution indicating the importance of MODIS LST as a proxy for Ta. These countrywide Ta grids provide vital tools for the planning of environmental and economic developments in the era of global climate change.

No abstract

Crops such as cannabis, poppy, and coca tree are used to make illicit and addictive drugs. Detection and mapping of such crops can be significant for the controlled growth of the plants, thus supporting the prevention of illegal production. Remote sensing has the ability to monitor areas for cannabis growing. However, in the scientific literature, there is relatively little information on the spectral features of cannabis. Here in this study, we aim to: (1) offer a literature review on the studies investigating Cannabis sativa L. using remote sensing data; (2) define the spectral features of cannabis fields and other plants found in areas where cannabis is produced in northern Turkey; (3) apply machine learning algorithms for distinguishing cannabis from non-cannabis fields. For the purposes of this study, high-resolution imagery from PlanetScope satellites was used. The investigation showed that the most significant difference between cannabis and the other investigated plants was noticed in May–June. The classification results showed that, with Random Forest (RF) cannabis, fields can be accurately classified with accuracy higher than 93%. Following these results, the investigations with machine learning techniques showed promising results for classifying cannabis fields.