Sample Questions for Students
1. Question
Write True or False
(a) Standardized anomalies can not be used for conveniently comparing two or more variables which have significantly different magnitudes (e.g. rainfall against temperature).
(b) Standardized Anomalies do not indicate how many times the value of a variable, say rainfall, is bigger or smaller than the Standard Deviation.
(c) Standard deviation generally indicates the overall amount of deviation of data values from the mean.
(d) Regression analysis is a statistical process for estimating the relationships among variables (i.e. dependent against independent variables).
(e) In order to correctly read terciles, the category with highest probability is mentioned first, followed by the category of the next highest probability
(f) In order to correctly read terciles, the category with lowest probability is mentioned first, followed by the category of the next lowest probability
2. Question
The medical doctor wants to see whether there is an increase in the number of skin disease cases when there is unusually high temperature during JFM season for the past thirty years; he/she wants to have a graph (i.e. time series) between these two parameters. What variables/statistic will you need to calculate in order to produce this graph?
3.
3.1 What statistic would you use to determine if there is a linear relationship between number of cold drinks sold during JJA and temperature?
3.2 What method would you use to determine the linear relationship between number of cold drinks sold during JJA and temperature?
3.3 What method would you use to determine the linear relationship between number of cold drinks sold during JJA against temperature and number of parties?
4. Question
MAM 2017 rainfall forecast indicates Above Normal rains over Mwanza region where there is a community which is living in a valley next to a big river. What would you advise this community regarding these rains? There is another community which do rice farming activities in this place but lives far up on high grounds; how would you advise this community regarding these rains?
Answer
The community living in valley next to a big river are advised to stay alert for weather and climate forecasts as issued by the National Meteorological Service. They are also advised to consider, relocating from the low lying areas to higher grounds when possible. Due to possible flood damage to crops and food stocks, it is also recommended they store excess food stuff in higher grounds or in elevated stores.
The community living on high grounds are advised to stay alert for weather and climate forecasts as issued by the National Meteorological Service. Due to possible flood damage to crops, it is also recommended they cultivate water log resistant crops as well as start harvesting as soon as the crops are ready for harvest. Excess food stuffs should be well stored in order to mitigate potential food scarcity.
Note: The answer need not be exactly like this; the important thing is an idea on DRR advisories
Some Concepts on Short Range Climate Prediction
Note:These concepts provide answers or hints for answering all questions above and even those questions which may come in your exam!!
By Elias J. Lipiki, Meteorologist, Numerical Weather Prediction Section, Central Forecast Office, Tanzania Meteorological Agency, 20-05-2017.
Season: A period of three (3) months
Climatology: A period of thirty (30) years
Accumulation: A sum of the values of weather/climate parameter for a specified period of time
Mean: An average of the values of weather/climate parameter for a specified period of time
Deviation/Anomaly: A difference between a value of climate parameter from its long-term (i.e. climatological) mean
Standard Deviation: A square root of average of squared deviations of a climate parameter; it is, informally, the measure of overall deviation for all the values of a climate parameter in a sample
Variance: Average of squared deviations of a climate parameter; it, informally, measures the overall spread (or variability) of all the values of a climate parameter in sample
Standardized Anomaly: Deviation divided by Standard Deviation; it, informally, measure how many times a deviation of a climate parameter is bigger or smaller than Standard Deviation of a sample from which it came from.
Predictand or Dependent/Explained Variable: A climatological parameter whose values depend on another/other parameter(s)
Predictor or Independent/Explanatory Variable: A climatological parameter whose values are independent of another/other parameter(s)
Correlation: A measure of relationship (dependence) between one climate parameter (a dependent variable) and another climate parameter (an independent variable).
Pearson’s Linear Correlation Coefficient (PLCC): A measure of linear relationship between a dependent and an independent variable
Regression: A statistical estimation of the relationship between a dependent and independent variable(s)
Linear Regression (LR): Estimation of the linear relationship between a dependent and independent variable(s)
Simple Linear Regression (SLR): A LR in which there is one dependent and one independent variable
Multiple Linear Regression (MLR): A LR in which there is one dependent and several (i.e. more than one) independent variable
Multiplicity: This is a problem whereby as the number of predictors grow, the probability of making wrong forecasts grows exponentially (this has been proven to be true), eventually reaching one. To solve this problem, few predictors must be used.
Multicollinearity: This is a problem whereby some of the predictors which relates (or correlates) to each other end up wrongly expressed in a regression equation, resulting in completely different (unanticipated) effects of those predictors on the predictand. To solve this problem, a trick which determines predictors which are not related must be used, and MLR cannot do this. So Principle Component Regression (PCR) comes in handy here.
Training Period: A period during which the relationship (i.e. the model or equation) between a dependent and independent variable is determined; informally, a period when a model is trained to estimate (to hindcast)
Validation (or cross-validation) Period: A period after Training Period during which the hindcasts from the model (or equation) are compared to the observed values of a climate parameter
Forecast: Estimation of an unknown future climate parameter
Hindcast: Estimation of a known past climate parameter
Coefficient of Determination (R-square): A measure of relationship between estimated (i.e. hindcasts) and actual (i.e. observed) values of a climate parameter