Publications

Chauhan, S; Jethoo, AS; Varshney, V (2023). Leveraging Aqua and Terra satellite data for improved diurnal land surface temperature prediction: a comparative LSTM-based approach. REMOTE SENSING LETTERS, 14(7), 733-742.

Abstract
Our study aimed to enhance the understanding of Land Surface Temperature (LST) dynamics and its impact on environment using remote sensing techniques. By developing LST prediction models based on MODIS data, we analysed both day and night LST patterns in Bharatpur, Rajasthan, India. Multiple evaluation metrics were employed to assess the accuracy of models and performance comprehensive assessment. Our findings revealed valuable insights into the correlation between diurnal LST at different times. Notably, night-time LST predictions exhibited a higher Pearson's correlation coefficient compared to day-time predictions, indicating a stronger linear relationship. The determination coefficient (R (2)) indicated a good fit between the predicted and test LST. Model selection was guided by the Akaike's Information Criteria (AIC) and Bayesian Information Criteria (BIC). The lower AIC and BIC values associated with night-time LST predictions suggested a better balance between model fit and complexity. Our study contributes to the advancement of remote sensing and climate research through a robust LST model-based prediction framework. The insights gained from our results have implications for various fields including environmental monitoring, urban planning and climate change studies. By improved understanding of LST dynamics, we can make informed decisions and develop effective strategies for sustainable development and resource management.

DOI:
10.1080/2150704X.2023.2234553

ISSN:
2150-7058