GI for Sustainability

GI & Sustainability I

5508.1 - A Spatial Analysis of Climate Change Effects on Maize Productivity in Kenya

Wednesday, July 5
1:30 PM - 1:50 PM
Location: Maryland C

Climate change has intensified the risk of catastrophic natural disasters all over the world. Though impacts of the changes are global, third world economies are more at risk, primarily because of their high dependence on natural resources, poverty, low capacity to adapt, lack of technological prowess and the existence of other significant environmental stress. Moreover, little or no information about the change and applicable mitigation and adaptation measures exacerbate the situation. Although agriculture remains the backbone of Kenya’s economy, the sector’s dependence on natural resources increases its vulnerability to the aggravating impacts of climate change and variability. Climate system variations that impact staple food crops like maize (Zea mays) ultimately threaten the food security of the nation. This study aimed to examine and correlate major environmental factors affecting maize productivity to facilitate informed planning and designing of climate change mitigation and adaptation measures in order to sustain maize farming in the changing and varying conditions.
In this research, maize yield was used as the depended variable. A number of models were created and tested on how effective they would use various environmental factors to predict changes in the dependent variable.
An exploratory data analysis was done on the variables against maize yields to find any observable preliminary relationships. This was followed by an Ordinary Least Squares regression with all variables against yields. Model diagnostic statistics were examined and a GWR model run to explore more regional variations in the relationships between the variables and the dependent variable. These were done in ArcGIS and R.
An exploratory data analysis revealed that elevation and precipitation have the greatest influence on maize yields at 22.4% and 55.1% respectively. OLS regression with all variables indicated redundancy between temperature range and elevation showing VIF higher than 7.5 and therefore, it can be inferred that the two variables have a similar impact on maize yields. This is true because elevation affects temperature since the higher one goes, the cooler it becomes. Owing to this, subsequent analyses would use temperature range in the place of elevation.
The OLS yielded an adjusted R squared of 73.6% with a residual standard error of 628.3 on 34 degrees of freedom and a p-value of 2.379e-09. Besides, elevation and precipitation were found to have statistically significant coefficients. An OLS with these two variables yielded an Adjusted R squared of 56.7% and an AICc of 703. This shows a strong relationship with the dependent variable. However, this model turned out with a statistically significant Jarque-Bera statistic meaning that it had biased predictions. It is worth noting that every OLS model was followed with an analysis of the spatial autocorrelation of its resultant residuals and all were random.
The GWR model resulted in an impressive statistic: an Adjusted R2 of 74.9% with a relatively lower AICc of 684.1. This shows a very strong and solid spatial relationship between the independent variables and maize yields. This model also had random residuals. T-values for the relationships were examined and revealed that elevation (and temperature by extension) have a greater impact on maize yields in the southeastern parts of Kenya. These regions border the Indian Ocean and are relatively warmer throughout the year. On the other hand, precipitation has the greatest influence in the northwestern parts which are classified as arid and semi-arid.
In conclusion, temperature and precipitation are the main factors determining maize yields in Kenya. To improve yields and reduce risk of food insecurity, decisions on optimal times for maize cropping should revolve more around these factors. Besides, measures to maintain these factors within optimal levels are highly encouraged.

Dan Wanyama

Graduate Student
The University of North Alabama

I'm a graduate student with interests in GIS, agriculture, climate change and sustainability. My major work right now revolves around analyzing the effect of climate change on maize yields in Kenya and developing a suitability model to predict optimal maize growing areas in the changing environmental times.


Send Email for Dan Wanyama

Horst Kremers

CODATA-Germany Chair, Berlin (Germany)


Send Email for Horst Kremers


5508.1 - A Spatial Analysis of Climate Change Effects on Maize Productivity in Kenya

Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for A Spatial Analysis of Climate Change Effects on Maize Productivity in Kenya