Table 1 Aggregated results across all protected areas and vegetation types

	Contrast 1 group A vs group B (PA vs spillover)			Contrast 2 group B vs group C (spillover vs independent)			Contrast 3 group A vs group C (PA vs independent)
Vegetation class	A ≠ B Positive	A ≠ B Negative	A = B None	B ≠ C Positive	B≠C Negative	B = C None	A ≠ C Positive	A ≠ C Negative	A = C None
X27: Woody closed (>65%)	609	482	498	738	1,690	1,267	2,157	1,942	2,633
(%)	38.33	30.33	31.34	19.97	45.74	34.29	32.04	28.85	39.11
X28: Woody open (40–65%)	1,495	655	561	1,699	1,439	617	4,103	1,720	964
(%)	55.15	24.16	20.69	45.25	38.32	16.43	60.45	25.34	14.2
X29: Woody open (15–40%)	1,462	459	888	1,129	1,362	1,264	3,249	1,394	2,152
(%)	52.05	16.34	31.61	30.07	36.27	33.66	47.81	20.52	31.67
X30: Woody sparse (4–15%)	146	103	656	240	150	3,321	663	328	5,765
(%)	16.13	11.38	72.49	6.47	4.04	89.49	9.81	4.85	85.33
X31: Woody scattered (1–4%)	64	29	379	67	90	3,473	227	241	6,202
(%)	13.59	6.16	80.47	1.85	2.48	95.67	3.4	3.61	92.98
X32: Herbaceous closed (>65%)	60	88	577	250	2	487	134	157	6,311
(%)	8.28	12.14	79.59	33.83	0.27	65.9	2.03	2.38	95.59
X33: Herbaceous open (40–65%)	626	920	890	1,467	1,307	976	1,715	2,628	2,442
(%)	25.7	37.77	36.54	39.12	34.85	26.03	25.28	38.73	35.99
X34: Herbaceous open (15–40%)	1,330	949	641	1,803	1,640	325	3,096	2,829	882
(%)	45.55	32.5	21.95	47.85	43.52	8.63	45.48	41.56	12.96
X35: Herbaceous sparse (4–15%)	667	797	972	1,151	1,828	789	2,346	2,796	1,665
(%)	27.38	32.72	39.9	30.55	48.51	20.94	34.46	41.08	24.46
X36: Herbaceous scattered (1–4%)	293	232	914	407	1,422	1,933	941	1,672	4,181
(%)	20.38	16.13	63.56	10.82	37.8	51.38	13.85	24.61	61.54

Results were aggregated for comparisons using sampling polygons inside the protected area (area A), in the 0–5 km spillover zone (area B) and beyond 50 km from the protected area (area C). For each vegetation type, the observed frequency of meaningful effects is given in the row labelled with the vegetation class. Effects are classed as ‘positive’, ‘negative’ or ‘none’. ‘None’ indicates that effect sizes were measured but fell between −0.1 and 0.1 and thus were not considered meaningful. Although the candidate sample of protected areas for inclusion was 12,513 in all cases, these reported frequencies and percentages do not include protected areas for which effect sizes could not be estimated. Sample sizes of protected areas differed for each contrast, owing to either an insufficient number of sampling polygons or the absence of a particular vegetation type. The row below each vegetation type expresses the frequency as a percentage of the total number of protected areas for which that particular contrast could be successfully run. For example, 1,589 protected areas were included in estimates of the class X27 (canopy >65%) for Contrast 1 and 3,695 were included in Contrast 2. The corresponding data for 1988 are presented in Supplementary Table 3.

Search