Wildlife or crop production: the dilemma of conservation and human livelihoods in Serengeti, Tanzania

Figures & data

Figure 1. Locations of Bunda, Serengeti (Mugumu) and Ngorongoro districts (Loliondo) are indicated by the gray triangles. The studied villages in the west and east of the Serengeti National Park (SNP) are also indicated.

Table 1. Items included in the calculation of possession score

Items	Description	Value	Points
Quality of houses	No house	None	–
	Poor	Low	1
	Good	Medium	2
	Very good	High	3
Livestock holdings^a	No livestock	None	–
	1–20	Low	1
	21–50	Medium	2
	51–200	High	3
	201 and above	Very high	4
Household equipments	No equipment	None	–
	Cell phone	Low	1
	Radio	Low	1
	Television	Medium	2
	Bicycle	Medium	2
	Motorised transport	High	3
Land	No land
	Uncultivated land	None	–
	Cultivated land (acres)	Low	1
	0–1	Low	1
	1–2	Medium	2
	2–4	High	3
	4 and above	Very high	4
Note: ^aCow equals four goats/sheep where the conversion was based on market prices (adult cow = USD 100).

Table 2. Percentage of farming activities, household crop sizes and crop-raiding by wildlife in villages outside the Serengeti National Park (SNP) in relation to location (districts, Western and Eastern Serengeti), immigration status, education level, employment and wealth in terms of number of livestock and wealth index

	Engagement in crop production (N =  477)		Final logistic model		Acreage of farm plots (N = 316)		Final logistic		Is crop-raiding by wildlife a problem? (N = 477)		Final logistic model
Variables	Cultivating	Not cultivating	χ^2	p	Acreage	Acreage	χ^2	p	Yes	No	χ^2	df	p
Districts
Bunda villages	96.7% (146)	3.3% (5)	90.72	<0.001	90.2% (92)	9.8% (10)	0.57	NS	82.1% (124)	17.9% (27)	17.27	2	<0.001
Serengeti villages	93.1% (148)	6.9% (11)			87% (100)	13% (15)			74.8% (119)	25.2% (40)
Ngorongoro villages	60.5% (101)	39.5% (66)			88.9% (88)	11.1% (11)			61.7% (103)	38.3% (64)
Immigration status
Yes	93.8% (167)	6.2% (11)	24.18	<0.001	84.2% (112)	15.8% (21)	43.99	=0.036
No	76.3% (228)	23.7% (71)			91.8% (168)	8.2% (15)
Education level
Not educated	71.5% (103)	28.5% (41)	18.44	<0.001					61.1% (88)	38.9% (56)	13.51	1	<0.001
Educated	87.7% (292)	12.3% (41)							77.5% (258)	22.5% (75)
Employment
Employed	88.7% (125)	11.3% (16)	48.01	=0.028
Not employed	80.4% (270)	19.6% (66)
Number of livestock owned
Less than 50 (few, <50)	91.1% (306)	8.9% (30)	54.50	<0.001					75.3% (253)	24.7% (83)	43.49	1	=0.037
50 or more (many, ≥50)	63.1% (89)	36.9% (52)							66.0% (93)	34.0% (48)
Wealth index
Poor (≤0.49)	81.3% (343)	18.7% (79)	6.02	=0.014	90.3% (242)	9.7% (26)	4.99	=0.025	70.6% (298)	29.4% (124)	6.77	1	=0.009
Rich (≥0.50)	94.5% (52)	5.5% (3)			79.2% (38)	20.8% (10)			87.3% (48)	12.7% (7)
Note: N, sample size; NS, not significant, differences tested with χ^2 tests.

Table 3. Crop-raiding by different wildlife outside the SNP in relation to districts and number of owned livestock

	Problematic wildlife			Final logistic model
Variables (N = 353)	Small/medium-sized wildlife	Elephants	Migratory species	χ^2	df	p
Districts
Bunda villages	40.5% (51)	58.3% (74)	1.6% (2)	47.80	4	<0.001
Serengeti villages	45.1% (55)	44.3% (54)	10.7% (13)
Ngorongoro villages	74.0% (77)	16.3% (17)	9.6% (10)
Number of livestock owned
Less than 50 (few, <50)	48.8% (127)	45.4% (118)	5.8% (15)	8.56	2	=0.014
50 or more (many, ≥50)	60.2% (56)	29.0% (27)	10.8% (10)
Note: N, sample size, differences tested with χ^2 tests.

Table 4. Other causes of crop loss in villages outside the SNP in relation to districts, immigration status, education level, number of livestock owned and wealth index

	Other factors influencing crop loss			Final logistic model
Variables (N = 169)	Unpredictable rainfall	Crop-raiding by livestock	Inadequate farming skills	χ^2	df	p
Districts
Bunda villages	44.8% (26)	0.00	55.2% (32)	70.45	4	<0.001
Serengeti villages	42.0% (21)	0.00	58.0% (29)
Ngorongoro villages	24.6% (15)	52.5% (32)	23.0% (14)
Immigration status
Yes	51.5% (35)	1.5% (1)	47.1% (32)	25.29	1	=0.001
No	26.7% (27)	30.7% (31)	42.6% (43)
Education level
Not educated	41.5% (22)	30.2% (16)	28.3% (15)	10.15	2	=0.006
Educated	34.5% (40)	13.8% (16)	51.7% (60)
Number of livestock owned
Less than 50 (few, <50)	38.5% (45)	12.0% (14)	49.6% (58)	12.39	2	=0.002
50 or more (many, ≥50)	32.7% (17)	34.6% (18)	32.7% (17)
Wealth index
Poor (≤0.49)	37.9% (55)	22.1% (32)	40.0% (58)	10.13	2	=0.006
Rich (≥0.50)	29.2% (7)	–	70.8% (17)
Note: N, sample size, differences tested with χ^2 tests.

Table 5. Regression analysis of the factors influencing crop production in villages surrounding the SNP

	Agriculture		Factors affecting crop production
	Crop production	Acreage of farm plots	Crop-raiding by wildlife	Problematic wildlife	Other factors affecting crop production
Independent variables	t-Value	t-Value	t-Value	t-Value	t-Value
Location (western, eastern)	90.01***	0.01	15.21**	2.82**	−1.68
Immigration status	24.18	4.39**	2.79	−0.24	2.26**
Gender	1.15*	0.73	1.29	−0.05	−1.24
Education level	18.44	0.13	13.52***	0.20	1.61
Employment status	4.80	0.07	0.26**	−0.402	1.08
Number of livestock	54.51**	0.00	4.35	0.93	−0.91
Wealth index	6.02	4.99**	6.77*	−1.72	2.03*
r ^2	0.312***	0.074***	0.094***	0.058***	0.111***
Notes: A negative coefficient (t) indicates reduced likelihood. See section ‘Methods’ for definitions of independent and dependent variables.
Significance of variable: *p < 0.05, **p < 0.01, ***p < 0.001.