Search

Frost is a major threat to crop productivity in the Kenyan highlands. With agriculture being central to the Kenyan economy, every effort needs to be taken to alleviate losses especially on high value crops like tea, the leading foreign exchange earner. Current frost mapping efforts by SERVIR, a joint initiative between National Aeronautics and Space Administration (NASA) and the U.S. Agency for International Development (USAID), and its hub institution in Eastern and Southern Africa, the Regional Center for Mapping of Resources for Development (RCMRD), utilizes Moderate Resolution Imaging Spectroradiometer (MODIS) derived Land Surface Temperature (LST) to probabilistically map areas that have been affected by frost. In this paper, we assessed the accuracy of these frost maps by testing the performance of MYD11A1 MODIS product in indicating areas affected by frost. MODIS derived LST values corresponding to frost and no frost observation locations and dates were reclassified according to 6 predetermined categories representing frost severity levels. The overall accuracy of each threshold category as LST cutoff separating frost and no frost affected areas was determined. An overall performance measure was then estimated using a Receiver Operating Characteristics curve (ROC). Overall accuracies of 67.3%–71.9% among the thresholds were obtained. An area under the ROC curve of 0.69 was obtained, indicating a poor performance of MODIS LST to distinguish frost from no frost areas. This shows that although MODIS derived LST can be used to identify frost-affected areas, it is not on its own sufficient in discriminating these areas with high levels of accuracy. Revision of temperature thresholds is recommended, in addition to improved characterization of frost occurrence in the region to include other factors that may be affecting frost occurrence. These results stand to better prepare the agricultural sector for damaging weather-related events. [ABSTRACT FROM AUTHOR]

• In Kenya, topographic features have a strong influence on frost occurrence. • East facing slopes are at greater risk of frost than west facing slope. • Frost probability increases with elevation and catchment flow path length. • Frost probability is higher on concave than convex landscape. • With projected weather data, a logistic model can forecast frost probability. Frequent frost occurrences in the Kenyan highlands have had devastating effects on agricultural productivity. With inadequate management systems to mitigate the impacts, farmers have often had to bear the burden of losses resulting from frost damage. While agriculture in Kenya remains dependent on weather and climate, the agricultural economy of Kenya continues to suffer, underscoring the need for building local knowledge as basis for development of early warning systems. The current paper attempts to delineate frost zones by statistically characterizing them based on known risk factors related to topography (elevation, convexity, aspect, upslope flow length) and Land Surface Temperature (LST) derived from Moderate Resolution Imaging Spectroradiometer (MODIS). Through binary logistic regression, a logistic regression model was developed utilizing observation data (frost occurrence and non-occurrence) as a binary dependent variable to estimate the probability of frost occurrence. Assuming a 0.5 probability cut-off threshold between frost occurrence and non-occurrence, an overall accuracy of 81% with area under Receiver Operating Characteristics (ROC) Curve of 0.88 was obtained. No evidence of lack of model fit was detected. This model outperforms the currently operational model that utilizes MODIS LST alone to detect frost zones in the Kenyan tea plantations. It provides an improved method for effective delineation of frost zones by incorporating local topographic characteristics. [ABSTRACT FROM AUTHOR]