Nepal_Calc_SampleSize_SampleDistribution_Template

	A	B	C	D	E	F	G	H	I	J
1	Update values for all yellow cells, the pixelcounts. Important results are in blue. pi values were estimated from results of the pilot study (Aryal et al. 2021), but could be adjusted. S(p) was chosen as 0.01, a generally accepted target standard error, but this could be adjusted.
2	Calculating Ideal Total Sample Size
3	Strata of Agreement Map		DEG	LOSS	GAIN	ComboChange	nonforest	stable forest
4	variable name	equation
5	pixel count	GEE output	2809258	326996	6140120	22526445	96047970	65990510
6	wi (Area Weight)	PixelCount / PixelCountSum	0.01449256693	0.001686926376	0.03167601554	0.1162107617	0.4954979692	0.3404357603
7	pi (Proportion of target (forest degradation) in stratum; estimated from 2021 Aryal paper)	defined	0.6	0.15	0.2	0.4	0.1	0.35
8	Si (Standard error for stratum)	sqrt(pi*(1-pi))	0.4898979486	0.3570714214	0.4	0.4898979486	0.3	0.4769696007
9	S(p) (Target standard error)	defined	0.01
10
11	numerator coefficients	wi*Si	0.007099878809	0.0006023531988	0.01267040622	0.05693141374	0.1486493908	0.1623775087
12	numerator	sum of numerator_coefficients	0.3883309514
13	denominator	S(p)	0.01
14	n (Target Sample Size)	(numerator/denominator)^2	1508.009278
15
16	Double check the equation used for [% distribution without small strata] is still only excluding the strata which are deemed to be small with the updated values (<3% is a decent threshold, or any that are below the desired sample size when proportionally distributed into points) In the inital sample design only values from 1DEG, nonforest, and forest were included in the sum. All others were considered small strata, requiring a minimum sample size instead of proportional sample size to be well represented. A minimum sample size of 60 was chosen, but this could be changed.
17	Sample Distribution
18	Strata of Agreement Map	equation	DEG	LOSS	GAIN	ComboChange	nonforest	stable forest
19	proportional % distribution	wi as a percent (*100)	1.45%	0.17%	3.17%	11.62%	49.55%	34.04%
20	proportional distribution of points	proportional % distibution * n [rounded to integers]	22	3	48	175	747	513
21	% distribution without small strata	excluding the smallest classes, % distribution or remaining pixel counts [row2value/(sum of larger row2values)]					59.27%	40.73%
22	minimum sample size	defined	60
23
24	point distribution with minimum sample size	Update this equation if calculating for more than 1DEG, forest, and nonforest (n-[MinimumSampleSize#SmallStrata])%DistributionWithoutSmall	60	60	60	60	752	516	total	1508
25	15% duplicate review points	One QA/QC option, selecting proportional 15% from strata for duplicate review	9	9	9	9	113	77	total	226
26	point distribution for single interpretation	total points distributed - 15% duplicate review points	51	51	51	51	639	439	total	1282
27									check sum for rounding error	1508
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100