Table 1.

Environmental characteristics of 12 flightless, alpine New Zealand grasshopper species modelled in this study. The most important predictor of species distribution (from nine variables) is shown for each species (EMwms). Predictor variable importance scores for all models and all species are in electronic supplementary material, S8. Range-change percentages are based on the loss/gain/stability of pixels between vector binary maps (electronic supplementary material, S9). Predicted range changes include all gained pixels (with dispersal), while range change with no dispersal includes only gain of currently occupied habitat fragments.

species	elevation range (m.a.s.l)	number of presence points	number of absence points	most important predictor variable	% range change from present (with dispersal)		% range change from present (no dispersal)
					RCP2.6	RCP8.5	RCP2.6	RCP8.5
A. crassicauda	975–1680	16	933	mean diurnal range	−43	208	−61	−11
A. tumidicauda	600–1830	49	900	precipitation of driest month	−74	−35	−74	−37
B. collinus	1000–2000	30	919	mean diurnal range	−95	−66	−96	−89
B. nivalis	450–2000	93	856	isothermality	59	134	−9	−7
P. dugdali	400–1160	29	920	precipitation seasonality	−75	−96	−82	−97
P. nitidus	600–1830	91	858	mean temp. of driest quarter	12	17	−61	−35
S. australis	285–2020	278	671	annual mean temp.	−74	−93	−75	−93
S. campestris	0–1550	100	849	soil	−45	−67	−45	−67
S. childi	160–420	41	908	precipitation of driest month	−47	154	−100	−89
S. minutus	500–1180	40	909	mean diurnal range	4	191	−80	−44
S. piliferus	725–1100	45	901	mean temp. of wettest quarter	−15	−58	−45	−77
S. villosus	1370–2130	23	926	annual mean temp.	168	−77	−20	−94