Multi-linear
Regression
Each district had a distinct
combination of climate variables for SPR and API that produced the largest R2
value for that set of predictor variables.
Below are displayed the group of climate variables for each district
that, together in a multi-linear regression, explained the most variability
(highest R2 values) in API and SPR by climatic variables alone.
|
|
|
|
|
|
|
|||||||||||||||
|
|
Temperature |
|
|
|
|
|
Precipitation |
|
|
R^2 |
p>F |
|||||||||
|
Coastal |
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||
|
Uttara Kannara |
Jan |
Feb |
March |
May |
Oct |
Lag Dec |
Lag Nov |
|
|
0.67 |
0.07 |
|||||||||
|
Dakshina Kannara |
Lag Dec |
|
|
|
|
|
July |
Lag Nov |
|
|
0.30 |
0.03 |
||||||||
North Interior |
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||
|
Raichur |
Jan |
Feb |
|
|
|
|
|
April |
|
|
|
0.53 |
0.002 |
|||||||
|
Gulbarga |
Jan |
Feb |
Lag Nov |
|
|
|
|
Nov |
Lag Nov |
|
|
0.50 |
0.009 |
|||||||
|
Bidar |
Jan |
Lag Nov |
|
|
|
|
|
June |
|
|
|
0.51 |
0.002 |
|||||||
|
Belgaum |
Jan |
April |
May |
Lag Nov |
|
|
Lag Nov |
|
|
0.66 |
0.001 |
|||||||||
|
Bijapur |
Jan |
April |
May |
|
|
|
|
|
|
|
|
0.36 |
0.02 |
|
||||||
|
Dharwad |
April |
May |
July |
Sept |
Dec |
Lag Nov |
Lag Dec |
April |
Lag Nov |
|
|
0.87 |
0.0001 |
|
||||||
|
South Interior |
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||
|
Bellary |
Jan |
April |
May |
July |
Feb |
Lag Nov |
Lag Dec |
Lag Nov |
|
|
0.84 |
0.0001 |
||||||||
|
Shimoga |
April |
May |
Lag Nov |
Lag Dec |
|
|
April |
Nov |
Lag Nov |
0.87 |
0.0001 |
|||||||||
|
Chitradurga and Hassan |
April |
May |
July |
Lag Dec |
|
|
April |
Lag Nov |
|
|
0.70 |
0.0007 |
||||||||
|
Kodagu |
May |
|
|
|
|
|
|
Oct |
Nov |
Lag Nov |
0.63 |
0.003 |
||||||||
Bangalore |
July |
|
|
|
|
|
|
April |
Lag Nov |
Feb |
|
0.47 |
0.008 |
|||||||
|
Mandya |
|
|
|
|
|
|
|
Apri |
Lag Nov |
Feb |
|
0.59 |
0.002 |
|||||||
|
Chikmagalur |
April |
July |
Lag Dec |
|
|
|
|
April |
July |
Oct |
Lag Nov |
0.77 |
0.0004 |
|||||||
|
Mysore |
May |
Lag Nov |
Lag Dec |
|
|
|
|
April |
Nov |
|
|
0.62 |
0.002 |
|||||||
|
Tumkur |
Lag Dec |
|
|
|
|
|
May |
Nov |
Lag Nov |
0.67 |
0.002 |
|||||||||
|
Kolar |
Feb |
|
|
|
|
|
|
April |
July |
Feb |
Lag Nov |
0.41 |
0.03 |
|||||||
|
|
||||||||||||||||||||
|
BEST REGRESSION FOR SPR BY DISTRICT IN KARNATAKA, INDIA |
|
|
|
|
||||||
|
|
Temperature |
|
|
|
Precipitation |
|
R^2 |
p>F |
||
|
COASTAL |
|
|
|
|
|
|
|
|
|
|
|
Uttara Kannara |
Jan |
Feb |
|
|
|
Lag Nov |
|
|
0.37 |
0.02 |
|
Dakshina Kannara |
Jan |
April |
|
|
|
Nov |
July |
|
0.39 |
0.02 |
|
NORTH
INTERIOR |
|
|
|
|
|
|
|
|
|
|
|
Raichur |
Jan |
Feb |
|
|
|
April |
|
|
0.55 |
0.001 |
|
Gulbarga |
Jan |
Lag Nov |
|
|
|
Nov |
|
|
0.46 |
0.005 |
|
Bidar |
Jan |
Sept |
Lag Dec |
|
|
|
|
|
0.53 |
0.002 |
|
Belgaum |
Jan |
May |
June |
|
|
Nov |
|
|
0.62 |
0.001 |
|
Bijapur |
Jan |
April |
May |
|
|
|
|
|
0.39 |
0.01 |
|
Dharwad |
Jan |
April |
May |
June |
Lag Dec |
Nov |
April |
|
0.72 |
0.001 |
|
SOUTH INTERIOR |
|
|
|
|
|
|
|
|
|
|
|
Bellary |
Jan |
April |
May |
July |
|
|
|
|
0.60 |
0.001 |
|
Shimoga |
Jan |
April |
May |
Lag Dec |
|
Nov |
Lag Nov |
|
0.87 |
0.00005 |
|
Chitradurga and Hassan |
Jan |
April |
May |
Lag Dec |
|
Nov |
|
|
0.71 |
0.0003 |
|
Kodagu |
April |
May |
|
|
|
Nov |
Lag Nov |
|
0.56 |
0.002 |
|
Bangalore |
April |
July |
|
|
|
Nov |
Lag Nov |
|
0.60 |
0.001 |
|
Mandya |
May |
|
|
|
|
Feb |
April |
Lag Nov |
0.76 |
0.00005 |
|
Chikmagalur |
May |
July |
Lad Dec |
|
|
Lag Nov |
|
|
0.69 |
0.0002 |
|
Mysore |
Lag Nov |
Lag Dec |
|
|
|
Nov |
April |
|
0.63 |
0.0006 |
|
Tumkur |
Dec |
Lag Dec |
|
|
|
Nov |
Lag Nov |
|
0.58 |
0.002 |
|
Kolar |
Feb |
|
|
|
|
Feb |
April |
July |
0.48 |
0.007 |
The
variability in API and SPR were predicted, from 30% to 87%, by a combination of
climate variables alone. For
example, 87% of the variation in API in Dharwad was predicted by the
multi-linear regression model with April mean temperature, May mean temperature,
July mean temperature, September mean temperature, December mean temperature,
November mean temperature lagged one year, December mean temperature lagged one
year, April precipitation, and November precipitation as independent variables.
There was variability between districts as to the amount of variation in
malaria prevalence rates that was predicted by climatic variables alone. In districts such as Dakshina Kannara, where only 30% of the
variability in API was predicted by the best combination of climate variables
for that malaria prevalence rate, other factors besides climate were likely to
be more influential in determining climatic factors. However, in the districts with a high % of variability in API
or SPR predicted by climate variables alone, non-climatic factors do not have as
large of an effect.
