Supplementary Information 2 for Mapping seasonal human mobility across Africa using mobile phone location history and geospatial data

This supplementary information page includes OLS and Bayesian regression results and model performance metrics for 50 country-level models (25 FR and 25 TO) followed by 10 region-level models (5 FR and 5 TO). All tables are numbered as FR/TO pairs identified alphanumerically by a/b, respectively. For each enumerated table, the main table (on top) lists fixed parameters and their covariate zonal statistics in the first two columnns. The next four columns are fitting results for the optimal OLS regression model, and the last two columns are the Bayesian posterior mean and 95% credible interval. Beneath the main table are performance metrics for the OLS regression in the first two columns and the Bayesian hyperparameter and performance metrics in the last three columns (see Supplemental Information 1 for details on metrics).

Country Level OLS and Bayesian Models

All tables of country-level model outputs are in alphabetical order by country.

S1a. OLS and Bayesian Regressions for the Angola FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0326	0.0192	1.70	0.091	0.0315	[-0.0494, 0.113]
Mean GDP Per Capita	mean*	−0.251	0.0565	−4.44	<0.001	−0.258	[-0.405, -0.111]
Total Runoff	sum*	−0.200	0.0421	−4.75	<0.001	−0.223	[-0.285, -0.161]
Urban Extent	sum*	−0.260	0.0591	−4.40	<0.001	−0.271	[-0.435, -0.108]
Accessibility	max*	−0.650	0.171	−3.81	<0.001	−0.707	[-1.16, -0.255]
Infact Mortality Rate	max	0.181	0.0241	7.49	<0.001	0.184	[0.118, 0.249]
No Primary Education, Female	max	−0.0131	0.00347	−3.79	<0.001	−0.0126	[-0.0218, -0.0035]
Precipitation	max	0.0119	0.00550	2.17	0.031	0.0147	[0.00739, 0.0219]
Secondary Education, Female	max	−0.0179	0.00340	−5.28	<0.001	−0.0188	[-0.0285, -0.00925]
Precipitation x Secondary Education, Female	max, max	0.00150	0.000601	2.49	0.013	0.00170	[0.000895, 0.00251]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.261	Precision (Gaussian)	8.95	[8.61, 9.31]
Adj.R.sqr	0.242	Precision (Spatial)	47.4	[25.8, 72.7]
Pred.R.sqr	0.218	Precision (Temporal)	1,480	[209, 4490]
Adj.R2.to.R2	0.927	phi	0.288	[0.0748, 0.683]
Pred.R2.to.R2	0.835	rho (Spatial)	−0.746	[-0.91, -0.362]
PRESS	47.4	rho (Temporal)	0.663	[-0.219, 0.975]
	—	DIC	249
	—	DIC, Saturated	265
	—	Marginal Likelihood	52.1

S1b. OLS and Bayesian Regressions for the Angola TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0899	0.0237	3.79	<0.001	0.105	[-0.181, 0.393]
Underweight Children	mean	−0.0392	0.00674	−5.82	<0.001	−0.0321	[-0.0507, -0.0133]
Skin Reservoir Content	sum*	−0.0675	0.0326	−2.07	0.039	−0.0681	[-0.124, -0.0126]
Surface Pressure	max	−0.0390	0.00597	−6.54	<0.001	−0.0327	[-0.0487, -0.0167]
No Primary Education, Female	max	0.00766	0.00382	2.00	0.046	0.00543	[-0.00318, 0.014]
No Primary Education, Male	max	0.0112	0.00330	3.40	<0.001	0.00834	[-0.000147, 0.0167]
Wind V (Northward)	mean	0.000100	0.0631	0.00159	0.999	−0.0640	[-0.15, 0.0223]
Accessibility	mean*	−0.0147	0.0941	−0.156	0.876	−0.0871	[-0.308, 0.133]
Wind V (Northward) x Accessibility	mean, mean*	−0.378	0.173	−2.19	0.029	−0.175	[-0.409, 0.0591]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.174	Precision (Gaussian)	8.78	[8.11, 9.47]
Adj.R.sqr	0.153	Precision (Spatial)	65.1	[16.6, 175]
Pred.R.sqr	0.128	Precision (Temporal)	69.1	[14, 198]
Adj.R2.to.R2	0.884	phi	0.283	[0.0153, 0.847]
Pred.R2.to.R2	0.739	rho (Spatial)	0.487	[-0.823, 0.985]
PRESS	48.1	rho (Temporal)	0.855	[0.438, 0.989]
	—	DIC	251
	—	DIC, Saturated	329
	—	Marginal Likelihood	38.7

S2a. OLS and Bayesian Regressions for the Benin FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.117	0.0137	8.54	<0.001	0.117	[0.0759, 0.157]
Infact Mortality Rate	mean	−0.0803	0.0175	−4.60	<0.001	−0.0806	[-0.111, -0.0506]
Secondary Education, Female	mean	0.111	0.0253	4.38	<0.001	0.111	[0.0628, 0.159]
Mean GDP Per Capita	mean*	−0.235	0.0740	−3.18	0.002	−0.235	[-0.366, -0.104]
Precipitation	mean	0.0146	0.00458	3.19	0.002	0.0145	[0.00806, 0.021]
Accessibility	max*	0.413	0.0608	6.79	<0.001	0.414	[0.313, 0.515]
Underweight Children	mean	−0.00805	0.00579	−1.39	0.165	−0.00813	[-0.0212, 0.00488]
Precipitation x Accessibility	mean, max*	0.0411	0.0182	2.26	0.025	0.0406	[0.0165, 0.0646]
Precipitation x Underweight Children	mean, mean	−0.00403	0.00167	−2.41	0.017	−0.00403	[-0.0063, -0.00176]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.232	Precision (Gaussian)	22.6	[20.7, 24.7]
Adj.R.sqr	0.207	Precision (Spatial)	1,550	[94.7, 8170]
Pred.R.sqr	0.181	Precision (Temporal)	6,730	[319, 37300]
Adj.R2.to.R2	0.890	phi	0.237	[0.00578, 0.83]
Pred.R2.to.R2	0.778	rho (Spatial)	0.532	[-0.824, 0.996]
PRESS	11.5	rho (Temporal)	0.667	[-0.467, 0.998]
	—	DIC	−58.0
	—	DIC, Saturated	3.30
	—	Marginal Likelihood	127

S2b. OLS and Bayesian Regressions for the Benin TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.112	0.0151	7.42	<0.001	0.112	[0.0553, 0.168]
Infact Mortality Rate	mean	−0.0260	0.0123	−2.11	0.036	−0.0267	[-0.0603, 0.00678]
Underweight Children	mean	−0.0154	0.00624	−2.47	0.014	−0.0160	[-0.0357, 0.00374]
Urban Extent	sum*	−0.0790	0.0397	−1.99	0.048	−0.0853	[-0.197, 0.0257]
Precipitation	mean	0.0106	0.00494	2.16	0.032	0.0106	[0.00386, 0.0174]
Accessibility	max*	0.210	0.0362	5.81	<0.001	0.213	[0.127, 0.299]
Precipitation x Accessibility	mean, max*	0.0392	0.0120	3.26	0.001	0.0411	[0.0246, 0.0575]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.190	Precision (Gaussian)	19.2	[17.5, 20.9]
Adj.R.sqr	0.170	Precision (Spatial)	173	[36.1, 463]
Pred.R.sqr	0.132	Precision (Temporal)	17,600	[2230, 62300]
Adj.R2.to.R2	0.896	phi	0.289	[0.0249, 0.848]
Pred.R2.to.R2	0.693	rho (Spatial)	0.651	[-0.573, 0.987]
PRESS	15.0	rho (Temporal)	0.492	[-0.885, 0.978]
	—	DIC	−13.3
	—	DIC, Saturated	42.3
	—	Marginal Likelihood	101

S3a. OLS and Bayesian Regressions for the Botswana FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0122	0.0382	−0.319	0.750	0.00267	[-0.122, 0.126]
Wind V (Northward)	mean	0.178	0.0423	4.20	<0.001	0.119	[0.0485, 0.189]
Surface Pressure	max	−0.108	0.0422	−2.55	0.011	−0.0821	[-0.16, -0.00431]
Wind U (Eastward)	max	−0.101	0.0391	−2.59	0.010	−0.0710	[-0.132, -0.01]
Underweight Children	mean	−0.0258	0.0109	−2.37	0.018	−0.0247	[-0.0499, 0.000402]
Mean GDP Per Capita	mean*	0.200	0.0685	2.92	0.004	0.168	[-0.0129, 0.348]
Accessibility	max*	0.220	0.0732	3.01	0.003	0.181	[0.0426, 0.318]
Underweight Children x Mean GDP Per Capita	mean, mean*	0.0750	0.0258	2.90	0.004	0.0597	[-0.00655, 0.125]
Underweight Children x Accessibility	mean, max*	0.0613	0.0141	4.35	<0.001	0.0542	[0.0226, 0.0858]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.133	Precision (Gaussian)	6.83	[6.23, 7.47]
Adj.R.sqr	0.105	Precision (Spatial)	125	[16.1, 494]
Pred.R.sqr	0.0680	Precision (Temporal)	1,700	[68.7, 9600]
Adj.R2.to.R2	0.794	phi	0.151	[0.00212, 0.701]
Pred.R2.to.R2	0.510	rho (Spatial)	0.305	[-0.625, 0.908]
PRESS	42.7	rho (Temporal)	0.628	[-0.643, 0.995]
	—	DIC	259
	—	DIC, Saturated	322
	—	Marginal Likelihood	19.2

S3b. OLS and Bayesian Regressions for the Botswana TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0850	0.0229	3.72	<0.001	0.0861	[-0.0272, 0.199]
Underweight Children	mean	−0.0141	0.00409	−3.44	<0.001	−0.0124	[-0.0278, 0.00287]
LAI High Vegetation	max	0.0453	0.0111	4.09	<0.001	0.0312	[-0.0154, 0.0765]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0690	Precision (Gaussian)	8.59	[7.81, 9.33]
Adj.R.sqr	0.0620	Precision (Spatial)	28.3	[10.1, 64.4]
Pred.R.sqr	0.0500	Precision (Temporal)	30,900	[1020, 174000]
Adj.R2.to.R2	0.909	phi	0.242	[0.00562, 0.887]
Pred.R2.to.R2	0.734	rho (Spatial)	0.336	[-0.889, 0.991]
PRESS	47.5	rho (Temporal)	0.596	[-0.71, 0.994]
	—	DIC	237
	—	DIC, Saturated	288
	—	Marginal Likelihood	44.8

S4a. OLS and Bayesian Regressions for the Burkina Faso FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0639	0.0240	2.66	0.008	0.0612	[-0.0594, 0.178]
LAI High Vegetation	max	−0.0609	0.0212	−2.87	0.004	−0.0586	[-0.13, 0.013]
No Primary Education, Female	max	0.0182	0.00862	2.12	0.035	0.0173	[-0.0133, 0.0478]
Infact Mortality Rate	mean	−0.347	0.0947	−3.66	<0.001	−0.326	[-0.65, -0.000637]
Accessibility	max*	0.775	0.235	3.31	0.001	0.770	[-0.032, 1.57]
Infact Mortality Rate x Accessibility	mean, max*	4.35	0.714	6.09	<0.001	4.36	[1.94, 6.77]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.160	Precision (Gaussian)	9.42	[8.51, 10.3]
Adj.R.sqr	0.145	Precision (Spatial)	45.0	[11.9, 122]
Pred.R.sqr	0.124	Precision (Temporal)	1,210	[69.9, 6680]
Adj.R2.to.R2	0.907	phi	0.213	[0.0144, 0.728]
Pred.R2.to.R2	0.775	rho (Spatial)	0.0976	[-0.963, 0.956]
PRESS	37.1	rho (Temporal)	0.776	[0.305, 0.958]
	—	DIC	199
	—	DIC, Saturated	195
	—	Marginal Likelihood	15.1

S4b. OLS and Bayesian Regressions for the Burkina Faso TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0877	0.0253	3.46	<0.001	0.0790	[-0.0932, 0.244]
2m Dewpoint Temperature	max	0.0164	0.00529	3.10	0.002	0.0145	[0.00695, 0.022]
Accessibility	mean*	−0.0181	0.238	−0.0760	0.940	−0.0317	[-0.595, 0.532]
Precipitation	sum*	−0.114	0.0448	−2.54	0.012	−0.0880	[-0.149, -0.0266]
No Primary Education, Female	mean	−0.0326	0.0174	−1.88	0.062	−0.0301	[-0.0705, 0.0103]
Urban Extent	sum*	0.00135	0.0631	0.0214	0.983	0.00979	[-0.148, 0.168]
Accessibility x Precipitation	mean, sum	0.632	0.124	5.08	<0.001	0.544	[0.38, 0.709]
No Primary Education, Female x Urban Extent	mean, sum*	−0.124	0.0262	−4.72	<0.001	−0.126	[-0.187, -0.065]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.250	Precision (Gaussian)	9.76	[8.88, 10.6]
Adj.R.sqr	0.229	Precision (Spatial)	97.8	[22.2, 246]
Pred.R.sqr	0.202	Precision (Temporal)	105	[24.9, 288]
Adj.R2.to.R2	0.918	phi	0.269	[0.0196, 0.792]
Pred.R2.to.R2	0.809	rho (Spatial)	0.480	[-0.782, 0.975]
PRESS	35.0	rho (Temporal)	0.825	[0.496, 0.971]
	—	DIC	173
	—	DIC, Saturated	205
	—	Marginal Likelihood	22.7

S5a. OLS and Bayesian Regressions for the Cameroon FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0419	0.0236	1.78	0.077	0.0419	[-0.00926, 0.0931]
Accessibility	mean*	−0.193	0.0814	−2.38	0.018	−0.195	[-0.331, -0.0587]
Mean GDP Per Capita	mean*	0.298	0.124	2.40	0.017	0.297	[0.0898, 0.503]
Urban Extent	prop*	0.00635	0.00139	4.59	<0.001	0.00635	[0.00423, 0.00846]
Mean GDP Per Capita x Urban Extent	mean, prop	0.0288	0.00735	3.91	<0.001	0.0288	[0.0175, 0.0401]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.112	Precision (Gaussian)	10.1	[9.17, 11.1]
Adj.R.sqr	0.0960	Precision (Spatial)	32,000	[184, 217000]
Pred.R.sqr	0.0730	Precision (Temporal)	5,130	[150, 23300]
Adj.R2.to.R2	0.850	phi	0.384	[0.0281, 0.873]
Pred.R2.to.R2	0.647	rho (Spatial)	0.0478	[-0.928, 0.868]
PRESS	22.1	rho (Temporal)	0.620	[-0.521, 0.99]
	—	DIC	121
	—	DIC, Saturated	123
	—	Marginal Likelihood	30.7

S5b. OLS and Bayesian Regressions for the Cameroon TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0743	0.0255	−2.92	0.004	−0.0785	[-0.168, 0.0104]
LAI High Vegetation	mean	−0.108	0.0378	−2.86	0.005	−0.105	[-0.168, -0.0423]
Precipitation	sum*	0.0904	0.0389	2.33	0.021	0.101	[0.0504, 0.153]
Wind V (Northward)	max	−0.0948	0.0355	−2.67	0.008	−0.121	[-0.173, -0.0683]
Underweight Children	mean	−0.0104	0.00391	−2.67	0.008	−0.0114	[-0.0233, 0.000122]
Urban Extent	sum*	0.0361	0.0382	0.943	0.347	0.0308	[-0.0707, 0.131]
Underweight Children x Urban Extent	mean, sum*	−0.0466	0.00920	−5.07	<0.001	−0.0494	[-0.0745, -0.0248]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.195	Precision (Gaussian)	9.79	[8.8, 10.8]
Adj.R.sqr	0.172	Precision (Spatial)	1,320	[32, 9700]
Pred.R.sqr	0.144	Precision (Temporal)	19,000	[1340, 66900]
Adj.R2.to.R2	0.881	phi	0.443	[0.0223, 0.988]
Pred.R2.to.R2	0.740	rho (Spatial)	0.168	[-0.939, 0.991]
PRESS	25.6	rho (Temporal)	0.681	[-0.498, 0.994]
	—	DIC	143
	—	DIC, Saturated	134
	—	Marginal Likelihood	17.3

S6a. OLS and Bayesian Regressions for the Côte d’Ivoire FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0259	0.0263	−0.984	0.326	−0.0433	[-0.222, 0.135]
Evaporation	mean	−0.299	0.0529	−5.64	<0.001	−0.417	[-0.527, -0.307]
Secondary Education, Male	mean	0.0296	0.0110	2.70	0.007	0.0569	[0.0133, 0.101]
Evaporation	sum*	1.39	0.189	7.34	<0.001	2.09	[1.59, 2.59]
Infact Mortality Rate	max	0.0700	0.0240	2.92	0.004	0.105	[0.00224, 0.209]
Underweight Children	mean	−0.0378	0.00726	−5.21	<0.001	−0.0454	[-0.0791, -0.0117]
Urban Extent	prop*	0.0523	0.00669	7.83	<0.001	0.0786	[0.0607, 0.0968]
Mean GDP Per Capita	mean*	−1.64	0.423	−3.87	<0.001	−3.18	[-4.52, -1.85]
Underweight Children x Urban Extent	mean, prop*	−0.0103	0.00195	−5.31	<0.001	−0.0187	[-0.0238, -0.0136]
Urban Extent x Mean GDP Per Capita	prop, mean	−0.339	0.0618	−5.49	<0.001	−0.617	[-0.767, -0.469]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.258	Precision (Gaussian)	6.80	[6.29, 7.32]
Adj.R.sqr	0.236	Precision (Spatial)	76.4	[8.02, 429]
Pred.R.sqr	0.206	Precision (Temporal)	250	[38.5, 955]
Adj.R2.to.R2	0.914	phi	0.102	[0.0223, 0.23]
Pred.R2.to.R2	0.799	rho (Spatial)	0.521	[-0.683, 0.993]
PRESS	58.9	rho (Temporal)	0.768	[0.317, 0.962]
	—	DIC	315
	—	DIC, Saturated	313
	—	Marginal Likelihood	−37.6

S6b. OLS and Bayesian Regressions for the Côte d’Ivoire TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.00356	0.0252	0.141	0.888	−0.0136	[-0.217, 0.185]
2m Surface Temperature	mean	−0.0371	0.0188	−1.97	0.050	−0.0363	[-0.073, 0.000942]
Surface Pressure	mean	−0.195	0.0474	−4.12	<0.001	−0.165	[-0.304, -0.0245]
No Primary Education, Male	mean	−0.0750	0.0283	−2.65	0.009	−0.0666	[-0.15, 0.0171]
Underweight Children	mean	−0.0325	0.0104	−3.12	0.002	−0.0266	[-0.0572, 0.00427]
Mean GDP Per Capita	mean*	0.781	0.263	2.98	0.003	0.812	[-0.0545, 1.68]
Urban Extent	sum*	0.227	0.0827	2.75	0.006	0.218	[-0.0647, 0.505]
No Primary Education, Female	max	0.0591	0.0158	3.75	<0.001	0.0571	[0.00972, 0.104]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.135	Precision (Gaussian)	6.73	[6.23, 7.25]
Adj.R.sqr	0.115	Precision (Spatial)	31.1	[8.98, 78.1]
Pred.R.sqr	0.0940	Precision (Temporal)	68.8	[17.3, 175]
Adj.R2.to.R2	0.853	phi	0.377	[0.0541, 0.87]
Pred.R2.to.R2	0.695	rho (Spatial)	0.221	[-0.972, 0.995]
PRESS	63.1	rho (Temporal)	0.841	[0.58, 0.97]
	—	DIC	315
	—	DIC, Saturated	317
	—	Marginal Likelihood	−31.5

S7a. OLS and Bayesian Regressions for the Egypt FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0615	0.00611	10.1	<0.001	0.0634	[-0.0523, 0.181]
No Primary Education, Female	max	−0.0808	0.0163	−4.96	<0.001	−0.0768	[-0.124, -0.0298]
No Primary Education, Male	max	0.0968	0.0194	4.98	<0.001	0.0865	[0.0304, 0.142]
2m Surface Temperature	max	0.00638	0.000968	6.59	<0.001	0.00588	[0.00371, 0.00799]
LAI High Vegetation	max	−0.0531	0.00414	−12.8	<0.001	−0.0513	[-0.063, -0.0394]
Secondary Education, Male	max	−0.0200	0.00198	−10.1	<0.001	−0.0203	[-0.0258, -0.0149]
2m Surface Temperature x LAI High Vegetation	max, max	−0.00455	0.000586	−7.76	<0.001	−0.00477	[-0.00558, -0.00397]
2m Surface Temperature x Secondary Education, Male	max, max	−0.00171	0.000269	−6.36	<0.001	−0.00170	[-0.00207, -0.00134]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.373	Precision (Gaussian)	54.2	[51.1, 57.1]
Adj.R.sqr	0.366	Precision (Spatial)	351	[140, 708]
Pred.R.sqr	0.349	Precision (Temporal)	456	[45.2, 1670]
Adj.R2.to.R2	0.981	phi	0.0805	[0.000432, 0.454]
Pred.R2.to.R2	0.936	rho (Spatial)	0.830	[-0.18, 0.998]
PRESS	14.8	rho (Temporal)	0.911	[0.646, 0.995]
	—	DIC	−642
	—	DIC, Saturated	−566
	—	Marginal Likelihood	569

S7b. OLS and Bayesian Regressions for the Egypt TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0629	0.00627	10.0	<0.001	0.0680	[0.00239, 0.133]
Secondary Education, Male	mean	0.00463	0.00213	2.18	0.030	0.00658	[-0.00194, 0.0151]
Mean GDP Per Capita	mean*	0.110	0.0402	2.74	0.006	0.179	[0.056, 0.303]
Urban Extent	sum*	0.0589	0.0217	2.71	0.007	0.0761	[-0.00128, 0.153]
2m Surface Temperature	max	0.00761	0.00144	5.30	<0.001	0.00612	[0.00237, 0.00977]
Secondary Education, Male	max	−0.0333	0.00401	−8.30	<0.001	−0.0358	[-0.0502, -0.0214]
Secondary Education, Female	max	0.00326	0.00263	1.24	0.215	−0.000255	[-0.00948, 0.00895]
Mean GDP Per Capita	max*	−0.0602	0.0230	−2.61	0.009	−0.0721	[-0.143, -0.00102]
LAI High Vegetation	max	−0.0503	0.00735	−6.85	<0.001	−0.0635	[-0.0888, -0.0382]
Surface Pressure	max	0.0167	0.0149	1.12	0.262	0.00785	[-0.0404, 0.0562]
2m Surface Temperature x Secondary Education, Male	max, max	−0.00139	0.000272	−5.09	<0.001	−0.00168	[-0.00205, -0.0013]
Secondary Education, Female x Mean GDP Per Capita	max, max*	−0.00356	0.00137	−2.59	0.010	−0.00372	[-0.00918, 0.00175]
LAI High Vegetation x Surface Pressure	max, max	0.0315	0.00566	5.57	<0.001	0.0684	[0.0562, 0.0806]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.366	Precision (Gaussian)	52.2	[49.4, 55.1]
Adj.R.sqr	0.354	Precision (Spatial)	1,690	[110, 7950]
Pred.R.sqr	0.331	Precision (Temporal)	467	[138, 1380]
Adj.R2.to.R2	0.966	phi	0.0438	[0.0116, 0.0968]
Pred.R2.to.R2	0.903	rho (Spatial)	0.868	[0.0318, 0.999]
PRESS	15.5	rho (Temporal)	0.692	[-0.337, 0.989]
	—	DIC	−624
	—	DIC, Saturated	−553
	—	Marginal Likelihood	544

S8a. OLS and Bayesian Regressions for the Gabon FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.290	0.0411	7.06	<0.001	0.290	[0.187, 0.393]
Surface Pressure	mean	0.175	0.0279	6.29	<0.001	0.173	[0.117, 0.229]
No Primary Education, Male	mean	−0.183	0.0355	−5.16	<0.001	−0.181	[-0.25, -0.112]
Urban Extent	sum*	0.299	0.114	2.63	0.010	0.306	[0.0592, 0.556]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.291	Precision (Gaussian)	5.23	[4.14, 6.21]
Adj.R.sqr	0.273	Precision (Spatial)	3,700	[14.7, 24600]
Pred.R.sqr	0.240	Precision (Temporal)	16,500	[479, 68500]
Adj.R2.to.R2	0.937	phi	0.435	[0.0218, 0.964]
Pred.R2.to.R2	0.825	rho (Spatial)	0.0487	[-0.845, 0.917]
PRESS	25.2	rho (Temporal)	0.115	[-0.975, 0.985]
	—	DIC	153
	—	DIC, Saturated	223
	—	Marginal Likelihood	13.7

S8b. OLS and Bayesian Regressions for the Gabon TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0685	0.0343	2.00	0.048	0.0751	[-0.0455, 0.197]
Precipitation	mean	0.0918	0.0239	3.84	<0.001	0.0816	[0.0512, 0.112]
Skin Reservoir Content	mean	−3.43	0.787	−4.36	<0.001	−2.98	[-3.97, -1.98]
2m Surface Temperature	max	0.0634	0.0659	0.962	0.338	0.0553	[-0.0455, 0.157]
VIIRS Night-time Lights	max*	−0.120	0.0426	−2.83	0.005	−0.112	[-0.192, -0.031]
2m Surface Temperature x VIIRS Night-time Lights	max, max*	0.329	0.0611	5.38	<0.001	0.296	[0.203, 0.39]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.248	Precision (Gaussian)	7.47	[6.26, 8.75]
Adj.R.sqr	0.221	Precision (Spatial)	238	[13.7, 1240]
Pred.R.sqr	0.175	Precision (Temporal)	21,100	[1590, 78500]
Adj.R2.to.R2	0.890	phi	0.378	[0.0248, 0.911]
Pred.R2.to.R2	0.706	rho (Spatial)	0.124	[-0.9, 0.977]
PRESS	23.1	rho (Temporal)	0.194	[-0.901, 0.932]
	—	DIC	139
	—	DIC, Saturated	175
	—	Marginal Likelihood	22.0

S9a. OLS and Bayesian Regressions for the Ghana FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0468	0.0109	4.30	<0.001	0.0361	[-0.0566, 0.124]
Accessibility	mean*	−0.327	0.0849	−3.85	<0.001	−0.328	[-0.549, -0.108]
No Primary Education, Female	mean	−0.0400	0.0109	−3.67	<0.001	−0.0393	[-0.0678, -0.0108]
Urban Extent	prop*	−0.00368	0.000980	−3.76	<0.001	−0.00364	[-0.00604, -0.00124]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0810	Precision (Gaussian)	45.3	[41.8, 48.8]
Adj.R.sqr	0.0680	Precision (Spatial)	201	[35.4, 676]
Pred.R.sqr	0.0420	Precision (Temporal)	888	[52.8, 4120]
Adj.R2.to.R2	0.837	phi	0.154	[0.00286, 0.719]
Pred.R2.to.R2	0.523	rho (Spatial)	0.755	[-0.362, 0.997]
PRESS	5.49	rho (Temporal)	0.895	[0.603, 0.99]
	—	DIC	−187
	—	DIC, Saturated	−167
	—	Marginal Likelihood	173

S9b. OLS and Bayesian Regressions for the Ghana TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0798	0.0142	5.63	<0.001	0.0764	[0.022, 0.129]
No Primary Education, Female	mean	−0.0192	0.0101	−1.91	0.058	−0.0189	[-0.0457, 0.00784]
Accessibility	mean*	−0.255	0.0868	−2.93	0.004	−0.271	[-0.499, -0.0455]
Secondary Education, Female	mean	−0.00504	0.00185	−2.72	0.007	−0.00578	[-0.012, 3e-04]
Accessibility x Secondary Education, Female	mean*, mean	0.0239	0.00824	2.90	0.004	0.0245	[0.00344, 0.0457]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0640	Precision (Gaussian)	41.7	[37.2, 46]
Adj.R.sqr	0.0470	Precision (Spatial)	932	[81.6, 4090]
Pred.R.sqr	0.0130	Precision (Temporal)	2,780	[234, 14300]
Adj.R2.to.R2	0.725	phi	0.259	[0.0116, 0.82]
Pred.R2.to.R2	0.197	rho (Spatial)	0.671	[-0.607, 0.991]
PRESS	6.03	rho (Temporal)	0.634	[0.306, 0.841]
	—	DIC	−173
	—	DIC, Saturated	−150
	—	Marginal Likelihood	169

S10a. OLS and Bayesian Regressions for the Kenya FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.142	0.0109	13.0	<0.001	0.124	[0.0171, 0.213]
School Holidays	days	0.00360	0.000703	5.13	<0.001	0.00351	[0.00246, 0.00455]
Wind U (Eastward)	max	−0.0784	0.0130	−6.01	<0.001	−0.0676	[-0.0866, -0.0486]
Wind V (Northward)	max	0.0351	0.00571	6.15	<0.001	0.0275	[0.0184, 0.0365]
Mean GDP Per Capita	max*	0.108	0.0223	4.84	<0.001	0.0965	[0.0586, 0.134]
LAI Low Vegetation	mean	0.0977	0.0254	3.84	<0.001	0.0967	[0.0503, 0.143]
2m Surface Temperature	max	−0.000107	0.00321	−0.0332	0.974	−0.000202	[-0.00599, 0.00558]
Evaporation	max	0.0501	0.0130	3.84	<0.001	0.0419	[0.0234, 0.0604]
Underweight Children	max	0.00308	0.00193	1.59	0.111	0.00315	[-0.000342, 0.00665]
Total Runoff	sum*	0.0731	0.0191	3.82	<0.001	0.0633	[0.0345, 0.0921]
Accessibility	max*	−0.0229	0.0528	−0.434	0.664	−0.0312	[-0.113, 0.0505]
Mean GDP Per Capita	mean*	−0.0450	0.0329	−1.37	0.172	−0.0248	[-0.0821, 0.0327]
Urban Extent	sum*	−0.163	0.0153	−10.6	<0.001	−0.164	[-0.193, -0.135]
Skin Reservoir Content	sum*	−0.118	0.0423	−2.79	0.005	−0.0851	[-0.145, -0.0248]
LAI Low Vegetation x 2m Surface Temperature	mean, max	0.0589	0.00633	9.31	<0.001	0.0529	[0.0431, 0.0626]
Evaporation x Underweight Children	max, max	−0.0148	0.00173	−8.55	<0.001	−0.0152	[-0.0178, -0.0126]
Underweight Children x Total Runoff	max, sum*	0.0179	0.00290	6.15	<0.001	0.0194	[0.0152, 0.0235]
Evaporation x Accessibility	max, max*	0.134	0.0263	5.09	<0.001	0.130	[0.0905, 0.169]
Mean GDP Per Capita x Urban Extent	mean, sum	0.270	0.0259	10.4	<0.001	0.267	[0.217, 0.316]
Accessibility x Skin Reservoir Content	max, sum	−0.419	0.0770	−5.45	<0.001	−0.420	[-0.535, -0.304]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.293	Precision (Gaussian)	21.1	[20.4, 21.8]
Adj.R.sqr	0.281	Precision (Spatial)	682	[314, 1230]
Pred.R.sqr	0.262	Precision (Temporal)	552	[36.2, 2240]
Adj.R2.to.R2	0.958	phi	0.0837	[0.00551, 0.348]
Pred.R2.to.R2	0.894	rho (Spatial)	0.990	[0.948, 0.999]
PRESS	56.9	rho (Temporal)	0.873	[0.329, 0.997]
	—	DIC	−191
	—	DIC, Saturated	−59.5
	—	Marginal Likelihood	505

S10b. OLS and Bayesian Regressions for the Kenya TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0902	0.00906	9.96	<0.001	0.0757	[0.0202, 0.122]
School Holidays	days	0.00332	0.000701	4.74	<0.001	0.00309	[0.00189, 0.00429]
No Primary Education, Female	mean	0.0138	0.00271	5.11	<0.001	0.0157	[0.0074, 0.0241]
Urban Extent	sum*	−0.115	0.0139	−8.23	<0.001	−0.112	[-0.156, -0.0677]
Secondary Education, Male	max	0.00491	0.000972	5.06	<0.001	0.00530	[0.00214, 0.00845]
Underweight Children	max	0.00533	0.00162	3.29	0.001	0.00554	[0.000586, 0.0105]
Mean GDP Per Capita	max*	0.0443	0.0137	3.23	0.001	0.0474	[0.00569, 0.0891]
2m Surface Temperature	max	0.00123	0.00312	0.394	0.694	0.000235	[-0.0079, 0.00832]
Evaporation	max	0.0311	0.00931	3.34	<0.001	0.0300	[0.0126, 0.0475]
LAI Low Vegetation	mean	0.0398	0.0243	1.64	0.101	0.0576	[-0.00739, 0.123]
Skin Reservoir Content	sum*	−0.0498	0.0236	−2.11	0.035	−0.0408	[-0.0883, 0.00658]
2m Surface Temperature x Evaporation	max, max	0.0119	0.00204	5.86	<0.001	0.00854	[0.00527, 0.0118]
LAI Low Vegetation x Skin Reservoir Content	mean, sum*	0.192	0.0450	4.26	<0.001	0.153	[0.0852, 0.221]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.198	Precision (Gaussian)	24.1	[23.2, 25]
Adj.R.sqr	0.189	Precision (Spatial)	299	[125, 622]
Pred.R.sqr	0.177	Precision (Temporal)	462	[119, 1100]
Adj.R2.to.R2	0.955	phi	0.172	[0.0036, 0.798]
Pred.R2.to.R2	0.896	rho (Spatial)	0.987	[0.941, 0.999]
PRESS	56.5	rho (Temporal)	0.524	[-0.186, 0.953]
	—	DIC	−325
	—	DIC, Saturated	−206
	—	Marginal Likelihood	560

S11a. OLS and Bayesian Regressions for the Libya FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0616	0.0569	−1.08	0.281	−0.0605	[-0.37, 0.248]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0	Precision (Gaussian)	2.27	[2, 2.55]
Adj.R.sqr	0	Precision (Spatial)	135	[8.24, 706]
Pred.R.sqr	−0.0120	Precision (Temporal)	42.9	[8.75, 136]
Adj.R2.to.R2	—	phi	0.131	[0.00381, 0.611]
Pred.R2.to.R2	−Inf	rho (Spatial)	0.273	[-0.97, 0.995]
PRESS	92.0	rho (Temporal)	0.784	[0.416, 0.95]
	—	DIC	354
	—	DIC, Saturated	333
	—	Marginal Likelihood	17.3

S11b. OLS and Bayesian Regressions for the Libya TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.205	0.0440	4.65	<0.001	0.203	[-0.0281, 0.433]
Secondary Education, Female	mean	−0.0892	0.0182	−4.91	<0.001	−0.0893	[-0.121, -0.0576]
Surface Pressure	max	−0.139	0.0336	−4.14	<0.001	−0.137	[-0.195, -0.0784]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.100	Precision (Gaussian)	2.47	[2.25, 2.7]
Adj.R.sqr	0.0920	Precision (Spatial)	140	[10.5, 674]
Pred.R.sqr	0.0800	Precision (Temporal)	65.1	[10.2, 228]
Adj.R2.to.R2	0.924	phi	0.121	[0.00325, 0.571]
Pred.R2.to.R2	0.800	rho (Spatial)	0.409	[-0.916, 0.995]
PRESS	113	rho (Temporal)	0.787	[0.371, 0.956]
	—	DIC	479
	—	DIC, Saturated	577
	—	Marginal Likelihood	−46.0

S12a. OLS and Bayesian Regressions for the Mali FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.102	0.0301	3.40	<0.001	0.102	[0.0298, 0.175]
Accessibility	mean*	−0.151	0.0622	−2.43	0.016	−0.151	[-0.266, -0.0363]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0340	Precision (Gaussian)	6.70	[5.97, 7.5]
Adj.R.sqr	0.0290	Precision (Spatial)	1,530	[29.6, 9880]
Pred.R.sqr	0.0130	Precision (Temporal)	22,500	[2000, 77800]
Adj.R2.to.R2	0.831	phi	0.383	[0.021, 0.922]
Pred.R2.to.R2	0.391	rho (Spatial)	0.0219	[-0.988, 0.99]
PRESS	25.7	rho (Temporal)	0.148	[-0.973, 0.99]
	—	DIC	163
	—	DIC, Saturated	197
	—	Marginal Likelihood	−1.61

S12b. OLS and Bayesian Regressions for the Mali TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0190	0.0285	0.668	0.505	0.0197	[-0.0823, 0.122]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0	Precision (Gaussian)	8.16	[7.27, 9.1]
Adj.R.sqr	0	Precision (Spatial)	127	[20.4, 461]
Pred.R.sqr	−0.0120	Precision (Temporal)	13,900	[646, 58800]
Adj.R2.to.R2	—	phi	0.357	[0.0221, 0.908]
Pred.R2.to.R2	−Inf	rho (Spatial)	0.375	[-0.945, 0.997]
PRESS	23.1	rho (Temporal)	0.412	[-0.878, 0.987]
	—	DIC	132
	—	DIC, Saturated	139
	—	Marginal Likelihood	12.0

S13a. OLS and Bayesian Regressions for the Morocco FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0690	0.0179	3.86	<0.001	0.0742	[0.0216, 0.127]
School Holidays	days	0.00520	0.00157	3.32	0.001	0.00508	[0.00276, 0.00741]
No Primary Education, Male	mean	0.398	0.0761	5.24	<0.001	0.384	[0.24, 0.528]
Secondary Education, Male	mean	0.120	0.0249	4.82	<0.001	0.118	[0.0742, 0.162]
Evaporation	sum*	0.200	0.0540	3.69	<0.001	0.188	[0.103, 0.273]
Urban Extent	prop*	−0.00378	0.000993	−3.81	<0.001	−0.00398	[-0.0062, -0.00176]
LAI Low Vegetation	max	0.229	0.0492	4.66	<0.001	0.207	[0.101, 0.312]
Secondary Education, Female	max	−0.0539	0.0141	−3.82	<0.001	−0.0550	[-0.0853, -0.0248]
Total Runoff	sum*	−0.216	0.0530	−4.08	<0.001	−0.214	[-0.29, -0.139]
Surface Pressure	max	−0.116	0.0194	−5.99	<0.001	−0.116	[-0.15, -0.0804]
Total Runoff x Surface Pressure	sum*, max	0.135	0.0372	3.64	<0.001	0.158	[0.106, 0.209]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.291	Precision (Gaussian)	13.9	[12.8, 15.2]
Adj.R.sqr	0.270	Precision (Spatial)	272	[35.3, 1060]
Pred.R.sqr	0.237	Precision (Temporal)	293	[55.7, 908]
Adj.R2.to.R2	0.925	phi	0.115	[0.00169, 0.609]
Pred.R2.to.R2	0.812	rho (Spatial)	0.817	[-0.383, 1]
PRESS	26.1	rho (Temporal)	−0.406	[-0.948, 0.568]
	—	DIC	77.7
	—	DIC, Saturated	99.8
	—	Marginal Likelihood	89.6

S13b. OLS and Bayesian Regressions for the Morocco TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0170	0.0147	1.15	0.250	0.0169	[-0.0165, 0.0502]
School Holidays	days	0.00802	0.00177	4.54	<0.001	0.00777	[0.00541, 0.0101]
Urban Extent	prop*	−0.000764	0.000729	−1.05	0.295	−0.000777	[-0.00216, 0.000606]
No Primary Education, Female	max	0.0266	0.0202	1.32	0.188	0.0264	[-0.0097, 0.0625]
2m Dewpoint Temperature	max	−0.00844	0.00498	−1.70	0.091	−0.00749	[-0.0143, -0.000661]
Infact Mortality Rate	max	0.0494	0.0116	4.26	<0.001	0.0491	[0.027, 0.0712]
Urban Extent x No Primary Education, Female	prop*, max	0.00389	0.00132	2.95	0.003	0.00397	[0.00162, 0.00632]
2m Dewpoint Temperature x Infact Mortality Rate	max, max	0.0105	0.00281	3.74	<0.001	0.0105	[0.00648, 0.0145]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.222	Precision (Gaussian)	15.0	[14, 16.1]
Adj.R.sqr	0.205	Precision (Spatial)	3,970	[2220, 8150]
Pred.R.sqr	0.181	Precision (Temporal)	6,840	[1110, 14300]
Adj.R2.to.R2	0.925	phi	0.444	[0.0218, 0.99]
Pred.R2.to.R2	0.815	rho (Spatial)	0.748	[-0.336, 0.996]
PRESS	23.6	rho (Temporal)	0.201	[-0.985, 0.994]
	—	DIC	56.4
	—	DIC, Saturated	58.8
	—	Marginal Likelihood	101

S14a. OLS and Bayesian Regressions for the Mozambique FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0401	0.0200	−2.01	0.046	−0.0465	[-0.138, 0.0458]
Infact Mortality Rate	mean	−0.0280	0.0135	−2.08	0.039	−0.0158	[-0.0634, 0.0329]
No Primary Education, Male	mean	0.0197	0.00493	4.00	<0.001	0.0193	[0.0045, 0.0341]
Accessibility	max*	−0.834	0.115	−7.26	<0.001	−0.846	[-1.27, -0.428]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.184	Precision (Gaussian)	12.2	[11.1, 13.4]
Adj.R.sqr	0.173	Precision (Spatial)	56.0	[13.1, 160]
Pred.R.sqr	0.157	Precision (Temporal)	28,900	[3410, 103000]
Adj.R2.to.R2	0.944	phi	0.340	[0.0135, 0.914]
Pred.R2.to.R2	0.854	rho (Spatial)	0.562	[-0.581, 0.982]
PRESS	23.4	rho (Temporal)	0.598	[-0.539, 0.987]
	—	DIC	99.2
	—	DIC, Saturated	79.9
	—	Marginal Likelihood	33.3

S14b. OLS and Bayesian Regressions for the Mozambique TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0489	0.0193	−2.53	0.012	−0.0566	[-0.13, 0.0172]
Infact Mortality Rate	mean	−0.0524	0.0150	−3.49	<0.001	−0.0349	[-0.0841, 0.0153]
No Primary Education, Male	mean	0.0273	0.00511	5.33	<0.001	0.0264	[0.0118, 0.0409]
Urban Extent	sum*	0.259	0.112	2.32	0.021	0.203	[-0.105, 0.508]
Accessibility	max*	−0.583	0.176	−3.32	0.001	−0.659	[-1.19, -0.129]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.234	Precision (Gaussian)	12.5	[11.4, 13.8]
Adj.R.sqr	0.221	Precision (Spatial)	92.8	[18.2, 284]
Pred.R.sqr	0.201	Precision (Temporal)	29,900	[3260, 106000]
Adj.R2.to.R2	0.944	phi	0.462	[0.0535, 0.915]
Pred.R2.to.R2	0.859	rho (Spatial)	0.254	[-0.937, 0.98]
PRESS	21.9	rho (Temporal)	0.167	[-0.908, 0.956]
	—	DIC	88.3
	—	DIC, Saturated	64.8
	—	Marginal Likelihood	38.2

S15a. OLS and Bayesian Regressions for the Namibia FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0461	0.00994	4.64	<0.001	0.0447	[-0.0336, 0.122]
2m Surface Temperature	mean	−0.00712	0.00290	−2.45	0.015	−0.00540	[-0.0102, -0.000427]
Surface Pressure	mean	0.0179	0.00533	3.35	<0.001	0.0183	[0.00118, 0.0355]
Infact Mortality Rate	mean	−0.0207	0.0103	−2.01	0.046	−0.0258	[-0.0585, 0.00647]
LAI High Vegetation	mean	−0.0609	0.0210	−2.90	0.004	−0.0804	[-0.156, -0.00875]
Total Runoff	sum*	−0.0528	0.0235	−2.25	0.025	−0.0580	[-0.0962, -0.0197]
LAI High Vegetation x Total Runoff	mean, sum*	0.0844	0.0247	3.42	<0.001	0.0857	[0.0516, 0.12]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.162	Precision (Gaussian)	41.8	[38.4, 45.4]
Adj.R.sqr	0.144	Precision (Spatial)	165	[44.1, 380]
Pred.R.sqr	0.120	Precision (Temporal)	964	[235, 3580]
Adj.R2.to.R2	0.889	phi	0.0692	[0.00124, 0.327]
Pred.R2.to.R2	0.741	rho (Spatial)	0.964	[0.831, 0.999]
PRESS	8.31	rho (Temporal)	0.585	[-0.303, 0.916]
	—	DIC	−229
	—	DIC, Saturated	−201
	—	Marginal Likelihood	227

S15b. OLS and Bayesian Regressions for the Namibia TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0626	0.0113	5.56	<0.001	0.0620	[0.00237, 0.122]
LAI High Vegetation	mean	−0.105	0.0265	−3.97	<0.001	−0.116	[-0.213, -0.0216]
Infact Mortality Rate	mean	−0.0838	0.0170	−4.92	<0.001	−0.0833	[-0.135, -0.0314]
Secondary Education, Female	mean	−0.00527	0.00200	−2.63	0.009	−0.00568	[-0.0124, 0.00101]
Mean GDP Per Capita	mean*	−0.131	0.0554	−2.37	0.019	−0.138	[-0.307, 0.0297]
Evaporation	sum*	−0.0893	0.0206	−4.33	<0.001	−0.0845	[-0.118, -0.0515]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.116	Precision (Gaussian)	32.5	[29.3, 35.7]
Adj.R.sqr	0.101	Precision (Spatial)	660	[131, 1830]
Pred.R.sqr	0.0770	Precision (Temporal)	2,510	[180, 11500]
Adj.R2.to.R2	0.866	phi	0.284	[0.00935, 0.832]
Pred.R2.to.R2	0.660	rho (Spatial)	0.919	[0.605, 0.996]
PRESS	10.8	rho (Temporal)	0.659	[-0.146, 0.967]
	—	DIC	−159
	—	DIC, Saturated	−123
	—	Marginal Likelihood	194

S16a. OLS and Bayesian Regressions for the Niger FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0350	0.0316	1.11	0.269	0.0345	[-0.0386, 0.107]
Infact Mortality Rate	mean	0.0730	0.0126	5.81	<0.001	0.0736	[0.0475, 0.0997]
No Primary Education, Male	mean	0.0240	0.00688	3.49	<0.001	0.0252	[0.0108, 0.0397]
Evaporation	sum*	0.0895	0.0403	2.22	0.028	0.101	[0.0386, 0.163]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.219	Precision (Gaussian)	6.55	[5.71, 7.42]
Adj.R.sqr	0.205	Precision (Spatial)	107,000	[53.1, 379000]
Pred.R.sqr	0.176	Precision (Temporal)	22,800	[1980, 83000]
Adj.R2.to.R2	0.935	phi	0.401	[0.0187, 0.953]
Pred.R2.to.R2	0.805	rho (Spatial)	−0.817	[-0.99, -0.165]
PRESS	29.0	rho (Temporal)	0.466	[-0.909, 0.996]
	—	DIC	177
	—	DIC, Saturated	189
	—	Marginal Likelihood	−18.7

S16b. OLS and Bayesian Regressions for the Niger TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0432	0.0321	1.35	0.179	0.0470	[-0.0849, 0.178]
No Primary Education, Female	mean	−0.0396	0.0161	−2.46	0.015	−0.0318	[-0.0836, 0.0207]
No Primary Education, Male	mean	0.0326	0.0107	3.05	0.003	0.0357	[0.00116, 0.0706]
Evaporation	sum*	0.100	0.0430	2.33	0.021	0.121	[0.0506, 0.191]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.104	Precision (Gaussian)	6.71	[5.92, 7.55]
Adj.R.sqr	0.0870	Precision (Spatial)	56.2	[8.75, 199]
Pred.R.sqr	0.0560	Precision (Temporal)	18,400	[872, 84700]
Adj.R2.to.R2	0.842	phi	0.345	[0.0201, 0.899]
Pred.R2.to.R2	0.543	rho (Spatial)	−0.187	[-0.983, 0.96]
PRESS	29.8	rho (Temporal)	0.646	[-0.461, 0.989]
	—	DIC	173
	—	DIC, Saturated	187
	—	Marginal Likelihood	−18.3

S17a. OLS and Bayesian Regressions for the Nigeria FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0582	0.00938	6.20	<0.001	0.0532	[-0.00432, 0.11]
Evaporation	mean	0.0504	0.0101	4.98	<0.001	0.0284	[0.0125, 0.0441]
Accessibility	mean*	−0.0701	0.0289	−2.42	0.016	−0.0773	[-0.164, 0.00871]
Underweight Children	mean	0.0127	0.00192	6.62	<0.001	0.0120	[0.00696, 0.017]
Urban Extent	sum*	−0.117	0.0254	−4.60	<0.001	−0.123	[-0.199, -0.0481]
Total Runoff	sum*	−0.00269	0.0134	−0.200	0.841	0.00527	[-0.0228, 0.0331]
Underweight Children	max	−0.00382	0.00133	−2.87	0.004	−0.00407	[-0.00775, -0.000397]
No Primary Education, Female	max	0.0106	0.00448	2.37	0.018	0.00984	[-0.00375, 0.0234]
Infact Mortality Rate	mean	−0.0214	0.00628	−3.41	<0.001	−0.0171	[-0.0343, 9.68e-05]
Secondary Education, Male	max	0.000815	0.00114	0.717	0.473	0.00161	[-0.00193, 0.00515]
Total Runoff x Underweight Children	sum*, max	−0.00393	0.00104	−3.79	<0.001	−0.00274	[-0.00434, -0.00114]
Underweight Children x No Primary Education, Female	max, max	−0.00211	0.000456	−4.64	<0.001	−0.00251	[-0.00388, -0.00115]
Infact Mortality Rate x Secondary Education, Male	mean, max	0.00339	0.000714	4.74	<0.001	0.00298	[0.000888, 0.00506]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.167	Precision (Gaussian)	40.9	[39, 42.7]
Adj.R.sqr	0.155	Precision (Spatial)	223	[106, 412]
Pred.R.sqr	0.141	Precision (Temporal)	390	[153, 832]
Adj.R2.to.R2	0.929	phi	0.272	[0.033, 0.71]
Pred.R2.to.R2	0.845	rho (Spatial)	0.454	[-0.895, 0.992]
PRESS	27.6	rho (Temporal)	0.526	[0.0889, 0.822]
	—	DIC	−690
	—	DIC, Saturated	−573
	—	Marginal Likelihood	650

S17b. OLS and Bayesian Regressions for the Nigeria TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0937	0.0126	7.45	<0.001	0.0854	[-0.0544, 0.211]
School Holidays	days	0.00315	0.000809	3.89	<0.001	0.00302	[0.00118, 0.00486]
Evaporation	mean	0.0190	0.00789	2.41	0.016	0.00759	[-0.00547, 0.0206]
Accessibility	mean*	−0.0643	0.0321	−2.00	0.046	−0.0473	[-0.154, 0.0592]
Urban Extent	sum*	−0.0875	0.0258	−3.39	<0.001	−0.0811	[-0.17, 0.00806]
LAI High Vegetation	max	0.0235	0.00749	3.13	0.002	0.0175	[-0.00511, 0.04]
Infact Mortality Rate	max	0.00872	0.00398	2.19	0.029	0.00932	[-0.00513, 0.0237]
Secondary Education, Male	mean	−0.00141	0.00109	−1.29	0.197	−0.000920	[-0.00465, 0.00281]
Underweight Children	max	0.000687	0.00116	0.590	0.555	−0.0000647	[-0.00425, 0.00408]
Secondary Education, Male x Underweight Children	mean, max	0.000283	0.000112	2.52	0.012	0.000259	[-0.000147, 0.000664]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0780	Precision (Gaussian)	38.0	[36.4, 39.7]
Adj.R.sqr	0.0680	Precision (Spatial)	125	[63.9, 222]
Pred.R.sqr	0.0560	Precision (Temporal)	153	[22.4, 460]
Adj.R2.to.R2	0.875	phi	0.107	[0.00152, 0.554]
Pred.R2.to.R2	0.721	rho (Spatial)	0.755	[-0.507, 0.997]
PRESS	30.2	rho (Temporal)	0.820	[0.437, 0.982]
	—	DIC	−639
	—	DIC, Saturated	−524
	—	Marginal Likelihood	618

S18a. OLS and Bayesian Regressions for the Rwanda FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0727	0.0196	3.71	<0.001	0.0722	[-0.00802, 0.152]
No Primary Education, Female	mean	−0.0528	0.0109	−4.85	<0.001	−0.0504	[-0.0853, -0.0135]
Urban Extent	prop*	−0.0122	0.00207	−5.90	<0.001	−0.0118	[-0.0183, -0.00483]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.294	Precision (Gaussian)	28.3	[23.8, 33.4]
Adj.R.sqr	0.279	Precision (Spatial)	738	[37.3, 3970]
Pred.R.sqr	0.241	Precision (Temporal)	29,400	[2130, 123000]
Adj.R2.to.R2	0.948	phi	0.409	[0.029, 0.919]
Pred.R2.to.R2	0.818	rho (Spatial)	0.338	[-0.956, 0.994]
PRESS	3.68	rho (Temporal)	0.583	[0.0191, 0.916]
	—	DIC	−39.2
	—	DIC, Saturated	−25.3
	—	Marginal Likelihood	66.8

S18b. OLS and Bayesian Regressions for the Rwanda TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0626	0.0225	2.78	0.007	0.0628	[-0.0146, 0.141]
Skin Reservoir Content	mean	−1.57	0.447	−3.51	<0.001	−1.45	[-2.07, -0.827]
Accessibility	mean*	0.213	0.0604	3.53	<0.001	0.218	[0.0404, 0.397]
2m Surface Temperature	max	−0.112	0.0328	−3.41	<0.001	−0.107	[-0.158, -0.0552]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.177	Precision (Gaussian)	21.3	[17.8, 25.5]
Adj.R.sqr	0.150	Precision (Spatial)	1,690	[17.5, 11300]
Pred.R.sqr	0.0940	Precision (Temporal)	27,300	[3620, 90200]
Adj.R2.to.R2	0.849	phi	0.461	[0.0396, 0.942]
Pred.R2.to.R2	0.528	rho (Spatial)	0.299	[-0.953, 0.994]
PRESS	4.94	rho (Temporal)	0.415	[-0.634, 0.964]
	—	DIC	−13.3
	—	DIC, Saturated	−1.28
	—	Marginal Likelihood	54.7

S19a. OLS and Bayesian Regressions for the Senegal FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.101	0.0175	5.78	<0.001	0.104	[0.0488, 0.158]
Accessibility	mean*	0.486	0.0818	5.94	<0.001	0.461	[0.277, 0.644]
Infact Mortality Rate	mean	−0.0914	0.0160	−5.71	<0.001	−0.0885	[-0.124, -0.0533]
Underweight Children	mean	0.00892	0.00290	3.08	0.002	0.00857	[0.00204, 0.0151]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.138	Precision (Gaussian)	11.9	[11, 12.9]
Adj.R.sqr	0.129	Precision (Spatial)	284	[43.1, 1040]
Pred.R.sqr	0.117	Precision (Temporal)	33,000	[4820, 108000]
Adj.R2.to.R2	0.934	phi	0.321	[0.0127, 0.897]
Pred.R2.to.R2	0.851	rho (Spatial)	0.322	[-0.943, 0.988]
PRESS	25.7	rho (Temporal)	0.771	[0.211, 0.977]
	—	DIC	117
	—	DIC, Saturated	175
	—	Marginal Likelihood	59.5

S19b. OLS and Bayesian Regressions for the Senegal TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.101	0.0187	5.39	<0.001	0.100	[0.031, 0.17]
Accessibility	mean*	0.490	0.0955	5.13	<0.001	0.470	[0.202, 0.738]
Infact Mortality Rate	max	−0.0619	0.0134	−4.61	<0.001	−0.0573	[-0.0998, -0.0144]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0870	Precision (Gaussian)	9.83	[9.15, 10.6]
Adj.R.sqr	0.0810	Precision (Spatial)	62.0	[25.3, 119]
Pred.R.sqr	0.0690	Precision (Temporal)	7,380	[721, 30400]
Adj.R2.to.R2	0.932	phi	0.158	[0.00261, 0.682]
Pred.R2.to.R2	0.795	rho (Spatial)	0.639	[-0.735, 0.998]
PRESS	34.4	rho (Temporal)	0.635	[-0.78, 0.998]
	—	DIC	184
	—	DIC, Saturated	247
	—	Marginal Likelihood	22.9

S20a. OLS and Bayesian Regressions for the South Africa FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0532	0.0164	3.24	0.001	0.0478	[0.00633, 0.089]
School Holidays	days	0.00558	0.00185	3.01	0.003	0.00534	[0.00275, 0.00792]
Total Runoff	max	−0.0440	0.0123	−3.58	<0.001	−0.0426	[-0.0612, -0.0241]
LAI High Vegetation	max	−0.104	0.0613	−1.70	0.090	−0.0862	[-0.208, 0.0346]
Wind U (Eastward)	max	−0.0171	0.0188	−0.905	0.367	−0.0243	[-0.0526, 0.00385]
LAI High Vegetation x Wind U (Eastward)	max, max	−0.310	0.0597	−5.18	<0.001	−0.249	[-0.345, -0.152]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.189	Precision (Gaussian)	32.7	[28.7, 37.1]
Adj.R.sqr	0.167	Precision (Spatial)	2,690	[182, 12100]
Pred.R.sqr	0.130	Precision (Temporal)	25,800	[2890, 91300]
Adj.R2.to.R2	0.884	phi	0.439	[0.0347, 0.94]
Pred.R2.to.R2	0.688	rho (Spatial)	0.675	[0.163, 0.936]
PRESS	6.87	rho (Temporal)	−0.0440	[-0.93, 0.914]
	—	DIC	−103
	—	DIC, Saturated	−100
	—	Marginal Likelihood	136

S20b. OLS and Bayesian Regressions for the South Africa TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0122	0.0152	−0.808	0.420	−0.0103	[-0.0731, 0.0529]
School Holidays	days	0.00705	0.00202	3.49	<0.001	0.00767	[0.00495, 0.0104]
Skin Reservoir Content	mean	−1.64	0.622	−2.64	0.009	−2.38	[-3.36, -1.4]
LAI Low Vegetation	max	0.0189	0.0429	0.440	0.661	0.0361	[-0.0726, 0.145]
Wind U (Eastward)	max	−0.0618	0.0171	−3.61	<0.001	−0.0677	[-0.0987, -0.0367]
LAI Low Vegetation x Wind U (Eastward)	max, max	−0.118	0.0327	−3.61	<0.001	−0.105	[-0.159, -0.0513]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.150	Precision (Gaussian)	31.4	[28.7, 34.2]
Adj.R.sqr	0.127	Precision (Spatial)	130	[49.6, 271]
Pred.R.sqr	0.0990	Precision (Temporal)	37,200	[10200, 102000]
Adj.R2.to.R2	0.847	phi	0.336	[0.0718, 0.741]
Pred.R2.to.R2	0.661	rho (Spatial)	0.728	[0.247, 0.939]
PRESS	7.64	rho (Temporal)	0.722	[-0.0568, 0.977]
	—	DIC	−96.2
	—	DIC, Saturated	−93.5
	—	Marginal Likelihood	134

S21a. OLS and Bayesian Regressions for the Tanzania FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0442	0.0118	3.73	<0.001	0.0267	[-0.137, 0.172]
Urban Extent	sum*	0.121	0.0220	5.52	<0.001	0.121	[0.0586, 0.183]
2m Dewpoint Temperature	max	−0.0278	0.00499	−5.57	<0.001	−0.0224	[-0.0326, -0.0122]
No Primary Education, Female	max	−0.00553	0.00184	−3.00	0.003	−0.00658	[-0.0117, -0.0015]
Underweight Children	max	0.00448	0.00170	2.64	0.008	0.00415	[-0.000214, 0.0085]
Surface Pressure	mean	0.0182	0.00296	6.16	<0.001	0.0177	[0.0107, 0.0247]
Mean GDP Per Capita	mean*	−0.0705	0.0327	−2.16	0.031	−0.0774	[-0.166, 0.0111]
Surface Pressure x Mean GDP Per Capita	mean, mean*	−0.0312	0.00619	−5.05	<0.001	−0.0307	[-0.0468, -0.0146]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.171	Precision (Gaussian)	16.1	[15.5, 16.8]
Adj.R.sqr	0.162	Precision (Spatial)	125	[70.2, 184]
Pred.R.sqr	0.151	Precision (Temporal)	124	[32.4, 293]
Adj.R2.to.R2	0.949	phi	0.112	[0.0217, 0.375]
Pred.R2.to.R2	0.881	rho (Spatial)	0.644	[-0.524, 0.982]
PRESS	54.4	rho (Temporal)	0.880	[0.727, 0.972]
	—	DIC	77.4
	—	DIC, Saturated	106
	—	Marginal Likelihood	228

S21b. OLS and Bayesian Regressions for the Tanzania TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.00574	0.0164	0.350	0.727	−0.0178	[-0.119, 0.0811]
Underweight Children	mean	0.00681	0.00149	4.58	<0.001	0.00668	[0.00278, 0.0106]
Urban Extent	sum*	0.253	0.0377	6.70	<0.001	0.272	[0.179, 0.365]
Urban Extent	prop*	−0.00376	0.00102	−3.69	<0.001	−0.00433	[-0.00686, -0.0018]
No Primary Education, Female	max	−0.00907	0.00201	−4.51	<0.001	−0.00949	[-0.0158, -0.00324]
2m Dewpoint Temperature	max	−0.0461	0.00752	−6.13	<0.001	−0.0348	[-0.0478, -0.0218]
Evaporation	max	0.0443	0.0136	3.25	0.001	0.0345	[0.0133, 0.0556]
Surface Pressure	mean	0.0149	0.00341	4.37	<0.001	0.0137	[0.00547, 0.0219]
2m Dewpoint Temperature x Evaporation	max, max	−0.0172	0.00473	−3.63	<0.001	−0.0127	[-0.0191, -0.00643]
2m Dewpoint Temperature x Surface Pressure	max, mean	0.00458	0.00153	3.00	0.003	0.00462	[0.00248, 0.00677]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.135	Precision (Gaussian)	15.4	[14.8, 16.2]
Adj.R.sqr	0.123	Precision (Spatial)	108	[38.2, 213]
Pred.R.sqr	0.109	Precision (Temporal)	218	[61.9, 580]
Adj.R2.to.R2	0.916	phi	0.0673	[0.0163, 0.154]
Pred.R2.to.R2	0.811	rho (Spatial)	0.658	[-0.663, 0.997]
PRESS	56.3	rho (Temporal)	0.801	[0.456, 0.952]
	—	DIC	112
	—	DIC, Saturated	130
	—	Marginal Likelihood	206

S22a. OLS and Bayesian Regressions for the Togo FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0571	0.0316	1.81	0.074	0.0590	[-0.059, 0.177]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0	Precision (Gaussian)	11.5	[9.58, 13.7]
Adj.R.sqr	0	Precision (Spatial)	166	[10.8, 794]
Pred.R.sqr	−0.0210	Precision (Temporal)	24,900	[1580, 111000]
Adj.R2.to.R2	—	phi	0.376	[0.038, 0.866]
Pred.R2.to.R2	−Inf	rho (Spatial)	−0.109	[-0.978, 0.972]
PRESS	9.32	rho (Temporal)	−0.0403	[-0.986, 0.984]
	—	DIC	47.5
	—	DIC, Saturated	58.5
	—	Marginal Likelihood	16.4

S22b. OLS and Bayesian Regressions for the Togo TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0784	0.0306	2.56	0.012	0.0795	[-0.015, 0.175]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0	Precision (Gaussian)	11.8	[9.86, 14.1]
Adj.R.sqr	0	Precision (Spatial)	2,290	[17, 15400]
Pred.R.sqr	−0.0210	Precision (Temporal)	36,600	[3310, 156000]
Adj.R2.to.R2	—	phi	0.376	[0.0179, 0.936]
Pred.R2.to.R2	−Inf	rho (Spatial)	−0.186	[-0.978, 0.943]
PRESS	8.74	rho (Temporal)	0.0490	[-0.984, 0.988]
	—	DIC	45.6
	—	DIC, Saturated	60.6
	—	Marginal Likelihood	19.5

S23a. OLS and Bayesian Regressions for the Uganda FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0913	0.0116	7.85	<0.001	0.0912	[-0.0159, 0.198]
2m Surface Temperature	mean	−0.0388	0.00740	−5.24	<0.001	−0.0344	[-0.0474, -0.0214]
Surface Pressure	mean	0.0163	0.00668	2.44	0.015	0.0135	[-0.00143, 0.0284]
Mean GDP Per Capita	mean*	0.0593	0.0227	2.61	0.009	0.0577	[-0.00766, 0.123]
Secondary Education, Female	max	0.00484	0.00148	3.26	0.001	0.00506	[0.000999, 0.00912]
Infact Mortality Rate	mean	0.0555	0.0117	4.72	<0.001	0.0556	[0.0233, 0.0878]
No Primary Education, Male	max	0.00975	0.00367	2.66	0.008	0.00961	[-0.000641, 0.0199]
Infact Mortality Rate x No Primary Education, Male	mean, max	−0.00504	0.00244	−2.06	0.039	−0.00554	[-0.0125, 0.00136]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0620	Precision (Gaussian)	8.96	[8.66, 9.26]
Adj.R.sqr	0.0570	Precision (Spatial)	49.4	[36.4, 66.1]
Pred.R.sqr	0.0500	Precision (Temporal)	307	[50.6, 952]
Adj.R2.to.R2	0.914	phi	0.121	[0.00435, 0.559]
Pred.R2.to.R2	0.799	rho (Spatial)	0.0408	[-0.985, 0.994]
PRESS	163	rho (Temporal)	0.863	[0.539, 0.986]
	—	DIC	855
	—	DIC, Saturated	1,050
	—	Marginal Likelihood	61.9

S23b. OLS and Bayesian Regressions for the Uganda TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0729	0.0165	4.41	<0.001	0.0737	[-0.033, 0.183]
Evaporation	mean	0.0415	0.0156	2.66	0.008	0.0365	[0.0109, 0.062]
Wind V (Northward)	mean	−0.0780	0.0179	−4.35	<0.001	−0.0592	[-0.0889, -0.0294]
No Primary Education, Female	mean	−0.0140	0.00484	−2.89	0.004	−0.0114	[-0.0226, -0.000241]
Urban Extent	prop*	−0.00419	0.00134	−3.13	0.002	−0.00370	[-0.00687, -0.000528]
Infact Mortality Rate	mean	0.132	0.0211	6.25	<0.001	0.128	[0.0774, 0.178]
Mean GDP Per Capita	mean*	−0.107	0.0336	−3.18	0.002	−0.102	[-0.19, -0.0141]
Underweight Children	mean	−0.0251	0.00527	−4.76	<0.001	−0.0249	[-0.0384, -0.0114]
Wind U (Eastward)	max	−0.0706	0.0334	−2.12	0.035	−0.0435	[-0.0974, 0.0103]
Surface Pressure	max	0.0642	0.0108	5.96	<0.001	0.0619	[0.0356, 0.0882]
Urban Extent	sum*	0.166	0.0292	5.70	<0.001	0.164	[0.099, 0.229]
Mean GDP Per Capita x Surface Pressure	mean*, max	−0.0569	0.0125	−4.57	<0.001	−0.0587	[-0.0906, -0.0268]
Underweight Children x Wind U (Eastward)	mean, max	0.0643	0.0142	4.53	<0.001	0.0411	[0.0216, 0.0606]
Infact Mortality Rate x Wind U (Eastward)	mean, max	−0.140	0.0490	−2.87	0.004	−0.108	[-0.176, -0.0408]
Underweight Children x Urban Extent	mean, sum*	0.0180	0.00499	3.61	<0.001	0.0191	[0.00684, 0.0314]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.135	Precision (Gaussian)	9.89	[9.56, 10.2]
Adj.R.sqr	0.123	Precision (Spatial)	57.2	[37.1, 86.1]
Pred.R.sqr	0.108	Precision (Temporal)	450	[61.7, 1630]
Adj.R2.to.R2	0.912	phi	0.108	[0.00288, 0.498]
Pred.R2.to.R2	0.799	rho (Spatial)	−0.0207	[-0.98, 0.987]
PRESS	137	rho (Temporal)	0.775	[0.153, 0.983]
	—	DIC	701
	—	DIC, Saturated	835
	—	Marginal Likelihood	140

S24a. OLS and Bayesian Regressions for the Zambia FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.131	0.0205	6.36	<0.001	0.131	[0.0464, 0.214]
Infact Mortality Rate	mean	−0.0483	0.0150	−3.21	0.002	−0.0483	[-0.0947, -0.00232]
No Primary Education, Male	mean	0.0846	0.0190	4.46	<0.001	0.0834	[0.0299, 0.137]
Mean GDP Per Capita	mean*	0.865	0.149	5.79	<0.001	0.846	[0.385, 1.3]
Skin Reservoir Content	max	−0.758	0.157	−4.84	<0.001	−0.800	[-1.02, -0.582]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.196	Precision (Gaussian)	12.1	[11, 13.3]
Adj.R.sqr	0.181	Precision (Spatial)	119	[18.4, 419]
Pred.R.sqr	0.157	Precision (Temporal)	23,100	[2090, 83300]
Adj.R2.to.R2	0.922	phi	0.327	[0.0153, 0.887]
Pred.R2.to.R2	0.799	rho (Spatial)	0.459	[-0.874, 0.991]
PRESS	20.1	rho (Temporal)	0.404	[-0.712, 0.975]
	—	DIC	86.7
	—	DIC, Saturated	143
	—	Marginal Likelihood	46.6

S24b. OLS and Bayesian Regressions for the Zambia TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.143	0.0222	6.43	<0.001	0.146	[0.0192, 0.271]
No Primary Education, Female	mean	0.0311	0.00737	4.22	<0.001	0.0474	[0.0033, 0.0933]
No Primary Education, Male	mean	−0.0534	0.0167	−3.19	0.002	−0.0726	[-0.148, 0.00289]
Urban Extent	prop*	−0.00279	0.00134	−2.09	0.038	−0.00384	[-0.00994, 0.00224]
Skin Reservoir Content	max	−0.700	0.173	−4.05	<0.001	−0.797	[-1.01, -0.586]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.134	Precision (Gaussian)	13.6	[12.3, 15.1]
Adj.R.sqr	0.118	Precision (Spatial)	28.0	[6.04, 72.9]
Pred.R.sqr	0.0960	Precision (Temporal)	16,300	[967, 74200]
Adj.R2.to.R2	0.878	phi	0.447	[0.0251, 0.964]
Pred.R2.to.R2	0.712	rho (Spatial)	0.458	[-0.899, 0.99]
PRESS	23.5	rho (Temporal)	0.338	[-0.91, 0.991]
	—	DIC	62.4
	—	DIC, Saturated	124
	—	Marginal Likelihood	55.5

S25a. OLS and Bayesian Regressions for the Zimbabwe FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0506	0.0200	2.53	0.012	0.0424	[-0.0763, 0.155]
Accessibility	mean*	0.372	0.0636	5.85	<0.001	0.383	[0.201, 0.566]
No Primary Education, Male	mean	−0.0836	0.0259	−3.23	0.001	−0.0884	[-0.157, -0.0199]
Underweight Children	mean	0.0539	0.0127	4.24	<0.001	0.0599	[0.0271, 0.0933]
LAI High Vegetation	max	−0.247	0.0537	−4.59	<0.001	−0.279	[-0.423, -0.139]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.248	Precision (Gaussian)	13.0	[11.7, 14.4]
Adj.R.sqr	0.234	Precision (Spatial)	184	[17.7, 767]
Pred.R.sqr	0.211	Precision (Temporal)	231	[36.9, 761]
Adj.R2.to.R2	0.943	phi	0.148	[0.00497, 0.629]
Pred.R2.to.R2	0.848	rho (Spatial)	0.503	[-0.861, 0.991]
PRESS	19.2	rho (Temporal)	0.704	[0.158, 0.954]
	—	DIC	77.0
	—	DIC, Saturated	98.9
	—	Marginal Likelihood	50.7

S25b. OLS and Bayesian Regressions for the Zimbabwe TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0255	0.0173	1.47	0.143	0.0245	[-0.0165, 0.0653]
Infact Mortality Rate	mean	−0.0572	0.00985	−5.80	<0.001	−0.0591	[-0.0794, -0.039]
Urban Extent	prop*	−0.00583	0.000698	−8.34	<0.001	−0.00590	[-0.00735, -0.00447]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.282	Precision (Gaussian)	16.1	[14.6, 17.6]
Adj.R.sqr	0.275	Precision (Spatial)	1,560	[711, 3760]
Pred.R.sqr	0.253	Precision (Temporal)	9,270	[467, 29400]
Adj.R2.to.R2	0.976	phi	0.400	[0.0204, 0.94]
Pred.R2.to.R2	0.898	rho (Spatial)	0.664	[-0.389, 0.992]
PRESS	14.4	rho (Temporal)	−0.0891	[-0.999, 0.985]
	—	DIC	23.5
	—	DIC, Saturated	34.6
	—	Marginal Likelihood	82.5

Regional Level OLS and Bayesian Models

All tables of regional-level model outputs are ordered north to south.

S26a. OLS and Bayesian Regressions for the Northern Region FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0547	0.0110	4.98	<0.001	−0.0933	[-0.345, 0.16]
2m Surface Temperature	max	0.0440	0.0109	4.05	<0.001	0.0447	[0.0199, 0.0693]
Mean GDP Per Capita	mean*	0.0205	0.0111	1.85	0.065	0.0230	[-0.0155, 0.0614]
Infact Mortality Rate	max	0.0238	0.0133	1.78	0.075	0.00768	[-0.0413, 0.0566]
Secondary Education, Female	mean	−0.00218	0.0140	−0.156	0.876	−0.0141	[-0.0824, 0.0542]
Mean GDP Per Capita x Infact Mortality Rate	mean*, max	−0.101	0.0143	−7.05	<0.001	−0.0866	[-0.13, -0.0428]
Mean GDP Per Capita x Secondary Education, Female	mean*, mean	−0.0743	0.0166	−4.47	<0.001	−0.0675	[-0.119, -0.0158]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0710	Precision (Gaussian)	13.2	[12.4, 14.2]
Adj.R.sqr	0.0660	Precision (Spatial)	2,130	[5.17, 14500]
Pred.R.sqr	0.0580	Precision (Temporal)	0.218	[0.00634, 0.803]
Adj.R2.to.R2	0.930	phi	341	[3.44, 2290]
Pred.R2.to.R2	0.826	rho (Spatial)	0.396	[-0.583, 0.952]
PRESS	141	rho (Temporal)	104	[37.7, 229]
	—	DIC	508
	—	DIC, Saturated	586
	—	Marginal Likelihood	531

S26b. OLS and Bayesian Regressions for the Northern Region TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0355	0.0119	2.99	0.003	0.0438	[-0.251, 0.337]
No Primary Education, Female	mean	−0.0160	0.0141	−1.14	0.256	−0.00110	[-0.0865, 0.113]
Precipitation	sum*	−0.0658	0.0149	−4.40	<0.001	−0.0446	[-0.0698, -0.0194]
Urban Extent	prop*	−0.0711	0.0129	−5.49	<0.001	−0.0627	[-0.098, -0.0273]
Surface Pressure	max	−0.0504	0.0114	−4.40	<0.001	−0.0408	[-0.0739, -0.00752]
VIIRS Night-time Lights	max*	0.0921	0.0118	7.80	<0.001	0.0531	[0.0265, 0.0794]
Infact Mortality Rate	max	0.00386	0.0122	0.316	0.752	−0.0101	[-0.0587, 0.0385]
No Primary Education, Female x Infact Mortality Rate	mean, max	0.0874	0.0133	6.58	<0.001	0.0790	[0.03, 0.128]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.122	Precision (Gaussian)	15.8	[14.3, 17.2]
Adj.R.sqr	0.117	Precision (Spatial)	1,620	[34.7, 10200]
Pred.R.sqr	0.108	Precision (Temporal)	0.188	[0.00315, 0.772]
Adj.R2.to.R2	0.958	phi	677	[4.62, 4570]
Pred.R2.to.R2	0.884	rho (Spatial)	0.642	[0.0133, 0.96]
PRESS	161	rho (Temporal)	22,300	[249, 141000]
	—	DIC	488
	—	DIC, Saturated	660
	—	Marginal Likelihood	388

S27a. OLS and Bayesian Regressions for the Central Region FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	−0.0945	0.0218	−4.34	<0.001	−0.125	[-0.328, 0.0885]
LAI Low Vegetation	mean	0.126	0.0196	6.43	<0.001	0.114	[0.0554, 0.173]
Infact Mortality Rate	mean	−0.250	0.0245	−10.2	<0.001	−0.212	[-0.294, -0.124]
No Primary Education, Male	mean	0.119	0.0220	5.42	<0.001	0.104	[0.0403, 0.168]
Urban Extent	sum*	−0.0702	0.0270	−2.60	0.010	−0.0899	[-0.168, -0.0118]
LAI Low Vegetation	max	−0.187	0.0250	−7.49	<0.001	−0.185	[-0.247, -0.124]
Accessibility	max*	0.118	0.0250	4.73	<0.001	0.106	[0.0359, 0.176]
Secondary Education, Female	max	0.0695	0.0227	3.07	0.002	0.0629	[-0.00217, 0.128]
Infact Mortality Rate x Urban Extent	mean, sum*	−0.167	0.0244	−6.84	<0.001	−0.182	[-0.253, -0.112]
No Primary Education, Male x Secondary Education, Female	mean, max	−0.293	0.0278	−10.5	<0.001	−0.300	[-0.381, -0.219]
LAI Low Vegetation x Urban Extent	mean, sum*	−0.131	0.0184	−7.10	<0.001	−0.134	[-0.184, -0.0855]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.226	Precision (Gaussian)	11.6	[10.7, 12.5]
Adj.R.sqr	0.214	Precision (Spatial)	14,300	[1320, 58300]
Pred.R.sqr	0.202	Precision (Temporal)	0.452	[0.0373, 0.922]
Adj.R2.to.R2	0.950	phi	1,330	[6.21, 8970]
Pred.R2.to.R2	0.895	rho (Spatial)	0.705	[-0.453, 0.993]
PRESS	99.8	rho (Temporal)	75,100	[217, 471000]
	—	DIC	439
	—	DIC, Saturated	527
	—	Marginal Likelihood	206

S27b. OLS and Bayesian Regressions for the Central Region TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0414	0.0176	2.36	0.019	0.0207	[-0.275, 0.317]
2m Surface Temperature	mean	0.217	0.0383	5.68	<0.001	0.150	[0.0871, 0.212]
VIIRS Night-time Lights	mean*	−0.134	0.0417	−3.21	0.001	−0.0128	[-0.0922, 0.0666]
No Primary Education, Male	mean	0.0659	0.0252	2.61	0.009	0.0705	[0.00143, 0.14]
Secondary Education, Female	mean	−0.208	0.0366	−5.67	<0.001	−0.191	[-0.295, -0.0877]
Urban Extent	sum*	0.0765	0.0214	3.57	<0.001	0.0732	[-0.00191, 0.149]
2m Surface Temperature	max	−0.184	0.0350	−5.27	<0.001	−0.134	[-0.191, -0.0766]
LAI High Vegetation	max	−0.0882	0.0181	−4.88	<0.001	−0.0722	[-0.115, -0.029]
Underweight Children	max	−0.276	0.0367	−7.52	<0.001	−0.210	[-0.305, -0.115]
Urban Extent x LAI High Vegetation	sum*, max	0.111	0.0233	4.75	<0.001	0.0764	[0.0311, 0.122]
No Primary Education, Male x LAI High Vegetation	mean, max	0.110	0.0249	4.43	<0.001	0.0653	[0.00694, 0.123]
VIIRS Night-time Lights x Underweight Children	mean*, max	−0.318	0.0683	−4.66	<0.001	−0.106	[-0.211, 0.000318]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.191	Precision (Gaussian)	10.2	[9.61, 10.8]
Adj.R.sqr	0.178	Precision (Spatial)	1,250	[4.57, 8180]
Pred.R.sqr	0.162	Precision (Temporal)	0.120	[0.00035, 0.692]
Adj.R2.to.R2	0.932	phi	349	[3.36, 2340]
Pred.R2.to.R2	0.847	rho (Spatial)	0.822	[0.308, 0.99]
PRESS	97.1	rho (Temporal)	148	[28.9, 477]
	—	DIC	481
	—	DIC, Saturated	588
	—	Marginal Likelihood	149

S28a. OLS and Bayesian Regressions for the Western Region FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0263	0.00728	3.61	<0.001	−0.450	[-2.36, 1.46]
School Holidays	days	0.0244	0.00617	3.95	<0.001	0.0174	[0.00372, 0.031]
Infact Mortality Rate	mean	0.0321	0.00815	3.94	<0.001	0.0294	[-0.0015, 0.0602]
Surface Pressure	max	0.0622	0.00780	7.98	<0.001	0.0546	[0.0247, 0.0845]
Accessibility	max*	−0.0235	0.00757	−3.10	0.002	−0.0176	[-0.0437, 0.00849]
VIIRS Night-time Lights	mean*	−0.0644	0.00867	−7.43	<0.001	−0.0441	[-0.0698, -0.0184]
Infact Mortality Rate x Surface Pressure	mean, max	−0.0547	0.00754	−7.26	<0.001	−0.0434	[-0.0695, -0.0172]
Surface Pressure x Accessibility	max, max*	0.0484	0.00722	6.71	<0.001	0.0430	[0.0201, 0.0658]
Surface Pressure x VIIRS Night-time Lights	max, mean*	0.0416	0.00803	5.18	<0.001	0.0319	[0.00875, 0.0551]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0550	Precision (Gaussian)	18.4	[17.7, 19]
Adj.R.sqr	0.0520	Precision (Spatial)	13,000	[1.95, 71700]
Pred.R.sqr	0.0450	Precision (Temporal)	0.179	[0.00225, 0.797]
Adj.R2.to.R2	0.948	phi	155	[23.9, 593]
Pred.R2.to.R2	0.829	rho (Spatial)	0.677	[-0.937, 1]
PRESS	265	rho (Temporal)	97.7	[49.3, 182]
	—	DIC	348
	—	DIC, Saturated	645
	—	Marginal Likelihood	1,020

S28b. OLS and Bayesian Regressions for the Western Region TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0503	0.00705	7.12	<0.001	−0.179	[-0.864, 0.502]
Evaporation	mean	0.0291	0.00702	4.14	<0.001	0.0118	[-0.00294, 0.0266]
No Primary Education, Female	mean	−0.0469	0.00919	−5.10	<0.001	−0.0207	[-0.0571, 0.0159]
Secondary Education, Female	mean	−0.0530	0.00977	−5.42	<0.001	−0.0391	[-0.0764, -0.00175]
Infact Mortality Rate	mean	0.00629	0.00750	0.838	0.402	0.0143	[-0.0146, 0.0434]
Surface Pressure	max	0.0512	0.00892	5.74	<0.001	0.0427	[0.0106, 0.0747]
Infact Mortality Rate x Surface Pressure	mean, max	−0.0215	0.00659	−3.27	0.001	−0.00981	[-0.0331, 0.0135]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0230	Precision (Gaussian)	21.0	[20.4, 470]
Adj.R.sqr	0.0210	Precision (Spatial)	863,000,000,000,000,000,000	[143, 3.02e+42]
Pred.R.sqr	0.0180	Precision (Temporal)	0.0783	[0.0487, 1]
Adj.R2.to.R2	0.902	phi	81.9	[60.5, 97.6]
Pred.R2.to.R2	0.772	rho (Spatial)	0.980	[0.913, 1]
PRESS	267	rho (Temporal)	85.2	[76.7, 5440000]
	—	DIC	142
	—	DIC, Saturated	468
	—	Marginal Likelihood	1,070

S29a. OLS and Bayesian Regressions for the Eastern Region FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0968	0.00701	13.8	<0.001	−0.361	[-1.41, 0.697]
School Holidays	days	0.0232	0.00516	4.49	<0.001	0.0199	[0.0122, 0.0275]
VIIRS Night-time Lights	mean*	0.0288	0.00832	3.47	<0.001	0.0322	[0.01, 0.0545]
No Primary Education, Female	mean	0.0721	0.0142	5.10	<0.001	0.0906	[0.0339, 0.147]
No Primary Education, Male	mean	−0.0899	0.0146	−6.14	<0.001	−0.0925	[-0.152, -0.0327]
Secondary Education, Female	mean	−0.0551	0.0152	−3.63	<0.001	−0.0492	[-0.0996, 0.0011]
Secondary Education, Female	max	0.0557	0.0121	4.61	<0.001	0.0492	[0.00578, 0.0926]
LAI High Vegetation	mean	0.0750	0.00889	8.44	<0.001	0.0571	[0.028, 0.0862]
Accessibility	max*	0.0266	0.00876	3.04	0.002	0.0362	[0.00508, 0.0673]
LAI High Vegetation	max	−0.0593	0.00849	−6.98	<0.001	−0.0394	[-0.0683, -0.0104]
Skin Reservoir Content	sum*	0.00709	0.00789	0.899	0.369	0.00130	[-0.0135, 0.0161]
LAI High Vegetation x Accessibility	mean, max*	0.0270	0.00501	5.39	<0.001	0.0228	[0.00469, 0.0408]
LAI High Vegetation x LAI High Vegetation	mean, max	−0.0162	0.00538	−3.01	0.003	−0.00713	[-0.0235, 0.00921]
Accessibility x Skin Reservoir Content	max, sum	−0.0491	0.00719	−6.83	<0.001	−0.0190	[-0.0314, -0.00668]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0680	Precision (Gaussian)	22.8	[22.2, 23.3]
Adj.R.sqr	0.0650	Precision (Spatial)	4,730	[0.556, 33400]
Pred.R.sqr	0.0610	Precision (Temporal)	0.141	[0.000861, 0.71]
Adj.R2.to.R2	0.953	phi	411	[17.7, 2370]
Pred.R2.to.R2	0.900	rho (Spatial)	0.647	[-0.951, 0.999]
PRESS	374	rho (Temporal)	148	[75.9, 279]
	—	DIC	−120
	—	DIC, Saturated	310
	—	Marginal Likelihood	1,800

S29b. OLS and Bayesian Regressions for the Eastern Region TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0671	0.00603	11.1	<0.001	−0.216	[-0.34, -0.0914]
School Holidays	days	0.0259	0.00501	5.16	<0.001	0.0142	[0.00647, 0.0219]
2m Dewpoint Temperature	mean	0.0559	0.00968	5.78	<0.001	0.0213	[0.00213, 0.0404]
2m Surface Temperature	mean	−0.0267	0.00813	−3.28	0.001	0.00349	[-0.0138, 0.0208]
LAI High Vegetation	mean	0.0395	0.00824	4.80	<0.001	0.0361	[0.0107, 0.0615]
VIIRS Night-time Lights	mean*	0.0567	0.0119	4.78	<0.001	0.0480	[0.0227, 0.0732]
Secondary Education, Female	max	0.0401	0.00631	6.35	<0.001	0.0407	[0.0148, 0.0666]
Precipitation	sum*	−0.0384	0.0111	−3.45	<0.001	−0.0288	[-0.0464, -0.0112]
Accessibility	max*	0.0581	0.00922	6.30	<0.001	0.0463	[0.0175, 0.075]
Total Runoff	mean	0.0283	0.0101	2.80	0.005	0.0185	[0.00523, 0.0319]
Urban Extent	sum*	0.00568	0.00923	0.615	0.538	0.00807	[-0.0192, 0.0353]
Urban Extent	prop*	−0.0474	0.0142	−3.34	<0.001	−0.0539	[-0.0888, -0.0189]
Precipitation	mean	−0.0196	0.0123	−1.59	0.113	−0.00821	[-0.0233, 0.00688]
LAI High Vegetation	max	−0.0348	0.00794	−4.39	<0.001	−0.0272	[-0.0537, -0.000617]
Precipitation x Accessibility	sum, max	−0.0483	0.00677	−7.14	<0.001	−0.0186	[-0.0294, -0.00774]
Total Runoff x Urban Extent	mean, sum*	−0.0774	0.0155	−5.01	<0.001	−0.0449	[-0.0639, -0.0259]
Total Runoff x Urban Extent	mean, prop*	0.0990	0.0243	4.08	<0.001	0.0518	[0.0239, 0.0796]
Urban Extent x Precipitation	sum*, mean	0.0507	0.0107	4.72	<0.001	0.0359	[0.0222, 0.0495]
Urban Extent x Precipitation	prop*, mean	−0.0516	0.0118	−4.36	<0.001	−0.0363	[-0.049, -0.0235]
Precipitation x Urban Extent	sum, sum	−0.0263	0.00650	−4.05	<0.001	−0.0165	[-0.029, -0.0041]
Precipitation x LAI High Vegetation	sum*, max	0.0207	0.00585	3.55	<0.001	0.000587	[-0.00831, 0.00948]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0790	Precision (Gaussian)	22.6	[22.1, 23]
Adj.R.sqr	0.0740	Precision (Spatial)	12,200	[113, 71600]
Pred.R.sqr	0.0680	Precision (Temporal)	0.252	[0.0101, 0.788]
Adj.R2.to.R2	0.938	phi	502	[21.4, 2700]
Pred.R2.to.R2	0.859	rho (Spatial)	−0.539	[-0.748, -0.21]
PRESS	345	rho (Temporal)	108	[56.9, 192]
	—	DIC	−203
	—	DIC, Saturated	127
	—	Marginal Likelihood	1,850

S30a. OLS and Bayesian Regressions for the Southern Region FR Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0701	0.0112	6.25	<0.001	−0.125	[-0.302, 0.053]
2m Dewpoint Temperature	mean	−0.0355	0.0127	−2.81	0.005	−0.0249	[-0.044, -0.00575]
Accessibility	mean*	−0.0656	0.0121	−5.43	<0.001	−0.0622	[-0.104, -0.02]
Urban Extent	sum*	−0.0658	0.0222	−2.96	0.003	−0.104	[-0.18, -0.029]
LAI Low Vegetation	max	0.0465	0.0142	3.28	0.001	0.0388	[-0.00369, 0.0808]
Secondary Education, Female	max	0.00988	0.0199	0.496	0.620	0.0253	[-0.0383, 0.0886]
LAI Low Vegetation x Secondary Education, Female	max, max	−0.0359	0.0120	−2.99	0.003	−0.0279	[-0.0638, 0.00817]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.0650	Precision (Gaussian)	21.9	[20.1, 23.9]
Adj.R.sqr	0.0570	Precision (Spatial)	3,240	[92, 20700]
Pred.R.sqr	0.0460	Precision (Temporal)	0.208	[0.00408, 0.783]
Adj.R2.to.R2	0.883	phi	731	[3.85, 5020]
Pred.R2.to.R2	0.715	rho (Spatial)	0.842	[-0.237, 1]
PRESS	65.0	rho (Temporal)	356,000,000	[747, 81300000]
	—	DIC	17.2
	—	DIC, Saturated	108
	—	Marginal Likelihood	465

S30b. OLS and Bayesian Regressions for the Southern Region TO Model
Parameters (Fixed)	Statistic1	OLS Regression Model				Bayesian Model
		Estimate	Error	t-Value	p-Value	Mean	95%CI
(Intercept)	—	0.0845	0.0106	7.99	<0.001	0.0675	[-0.119, 0.253]
Accessibility	mean*	−0.0469	0.0121	−3.88	<0.001	−0.0773	[-0.117, -0.0373]
Infact Mortality Rate	mean	0.0779	0.0136	5.73	<0.001	0.109	[0.061, 0.156]
Mean GDP Per Capita	max*	0.0373	0.0139	2.68	0.007	0.0431	[-0.00239, 0.0885]
LAI High Vegetation	mean	0.0521	0.0110	4.75	<0.001	0.0521	[0.0147, 0.0891]
Urban Extent	sum*	0.0143	0.0155	0.919	0.358	0.0277	[-0.0227, 0.0781]
LAI High Vegetation x Urban Extent	mean, sum*	−0.0827	0.00944	−8.76	<0.001	−0.0706	[-0.101, -0.0395]
1 Statistics denoted by asterisk (*) are log10 transformed zonal statistics.

OLS Performance		Bayesian Hyperparameters and Performance
Metric	Value	Hyperparameter	Mean	95%CI
R.squared	0.129	Precision (Gaussian)	17.7	[16.7, 18.6]
Adj.R.sqr	0.123	Precision (Spatial)	2,460	[23.7, 16500]
Pred.R.sqr	0.112	Precision (Temporal)	0.208	[0.00386, 0.823]
Adj.R2.to.R2	0.948	phi	205	[3.73, 1250]
Pred.R2.to.R2	0.863	rho (Spatial)	0.722	[-0.0131, 0.983]
PRESS	62.9	rho (Temporal)	385	[69.8, 1450]
	—	DIC	91.0
	—	DIC, Saturated	181
	—	Marginal Likelihood	498