Systematic errors in temperature estimates from MODIS data covering the western Palearctic and their impact on a...

Area of Science:

• Environmental Science
• Parasitology
• Remote Sensing

Background:

• Climate change is altering infectious disease patterns, necessitating accurate parasite habitat suitability models.
• Accurate, high-resolution temperature data are crucial for these models, influencing parasite development rates and distribution.
• Remote sensing, particularly Moderate Resolution Imaging Spectroradiometer (MODIS) sensors, offers potential for obtaining such temperature data.

Purpose of the Study:

• To comparatively analyze MODIS-derived surface temperatures against ground records in the western Palaearctic.
• To assess the accuracy of MODIS temperature data for parasite habitat suitability modeling.
• To quantify the error propagation of temperature uncertainties in parasite lifecycle models.

Main Methods:

• Comparative analysis of MODIS (Aqua and Terra) derived temperatures with ground-based surface temperature records.
• Evaluation of temperature estimation accuracy across different latitudes and seasons.
• Estimation of error propagation in parasite habitat suitability models using published examples.

Main Results:

• MODIS data showed overestimations of maximum temperatures and underestimations of minimum temperatures by 5-6 °C.
• Combined Aqua and Terra data improved accuracy around 35-44° N, with seasonal biases (overestimation in spring-summer, underestimation in autumn-winter).
• Temperature errors significantly impacted parasite habitat suitability models, with error propagation reaching up to 36%.

Conclusions:

• MODIS temperature data exhibit notable inaccuracies when compared to ground truth, particularly concerning extreme values and seasonal variations.
• Technical limitations of satellite sensors and orbital coverage contribute to temperature estimation errors.
• Careful image processing and awareness of MODIS data limitations are essential for reliable parasite habitat suitability modeling and disease distribution predictions.