sites sampled by year a - colorado state...
TRANSCRIPT
174
Table 2.1. Temporal and spatial replication of small mammal abundance and microhabitatsamples in the Sacramento Mountains, New Mexico (1991–1996). Asterisks indicate sitessampled also during fall of 1993 and winter and spring of 1994.
Sites Sampled By Year a Habitat 1991 1992 1993 1994 1995 1996
Mid-seral BLUF (0) DELW (27) DELW (60)* DELW (34) DELW (0) DELW (0)Mesic Forest GRES (0) HAYC (1) HAYC (120)* HAYC (0) HAYC (0)
ZINK (0) ZINK (90) ZINK (30) ZINK (0)WAYL (5) WAYL (51) WAYL (65)WIWH (27) WIWH (80) WIWH (14)LWIL (38) LWIL (83)
Replicates: 2 6 6 5 3 1
Late Seral GDSL (121) GDSL (0) GDSL (0)Mesic Forest SUNS (121) SUNS (0)Replicates: 0 0 0 2 2 1
Montane DELW (22) DELW (29)* DELW (54) DELW (0) DELW (0)Meadow HAYC (30) HAYC (35)* HAYC (40) HAYC (0)
WILL (30) WILL (0) WILL (74) WILL (0)LWIL (27) LWIL (0) LWIL (78)WAYL (19) WAYL (32) WAYL (54)WIWH (28) WIWH (0) WIWH (76)
Replicates: 0 6 6 6 3 1
Xeric Forest FIRE1 (0) JAMS (19) JAMS (94)* JAMS (18) JAMS (0) JAMS (0)SIXT (0) CARR (30) CARR (19) CARR (71) CARR (0)
DRYB (31) DRYB (40) DRYB (48) BRYB (0)FIRE2 (22) FIRE2 (99)* FIRE2 (0)TRAL (3) TRAL (118) TRAL (0)WALK (8) WALK (102) WALK (11) WALK (0)
Replicates: 2 6 6 6 4 1 a Number of trap stations where environmental variables were sampled is shown in parentheses. Site names anduniversal transverse mercator coordinates (zone 13) for center of trapping grid: BLUF–Bluff Springs (4303E, 36321N);CARR–Carr Gap (4547E, 36438N); DELW–Delworth (MF:4309E, 36330N; MM:4305E, 36324N); DRYB–Dry Burnt(4473E, 36458N); FIRE1–Fire 1(4552E, 36450N); FIRE2–Fire 2 (4547E, 36446N); GDSL–Goodsell (4322E, 36411N);GRES–Greasy Canyon (4418E 36328N); HAYC–Hay Canyon (MF:4396E, 36295N; MM:4393E, 36298N);JAMS–James Canyon (4447E, 36474N); LWIL–Lower Wills Canyon (MF:4338E, 36311N; MM:4331E, 36311N);SIXT–Sixteen Springs (4541E, 36491N); SUNS–Sunspot (4239E, 36288N); TRAL–Trail Canyon (4444E, 36215N);WALK–Walker (4521E, 36474N); WAYL–Wayland (MF:4375E, 36237N; MM:4375E, 36239N); WILL–Upper WillsCanyon(4304E, 36293N); WIWH–Willie White Canyon (MF:4322E, 36324N; MM:4313E, 36323N); ZINK–Zink Trick(4372E, 36513N).
175
Table 2.2. Attributes sampled at live-trapping stations to characterize habitat types and quantify microhabitat associations of five rodent species occurring in the SacramentoMountains, New Mexico.
SamplingVariable Method a Description
SLOPE PT average between 2 slope (%) measurements using a clinometer ELEV PT elevation to nearest 10 ft (converted to m) taken with an altimeter
HERBCAV PI % herbaceous cover averaged over 10 point-intercepts SHRBCAV PI % shrub cover averaged over 10 point-intercepts TREECAV PI % tree cover averaged over 10 point-intercepts ROCKCAV PI % rock cover averaged over 10 point-intercepts BARECAV PI % bare ground averaged over 10 point-intercepts LITDPAV PI litter depth (cm) averaged over 10 point-intercepts GFHTAV PI maximum grass/forb height (cm) averaged over 10 point-intercepts
SLOG1D and SLOG1V FP density (no./ha) and volume (m3) of stage-1 logs (mid-diam. 10–30 cm) SLOG2D and SLOG2V FP density (no./ha) and volume (m3) of stage-2 logs (mid-diam. 10–30 cm) SLOG3D and SLOG3V FP density (no./ha) and volume (m3) of stage-3 logs (mid-diam. 10–30 cm) SLOG4D and SLOG4V FP density (no./ha) and volume (m3) of stage-4 logs (mid-diam. 10–30 cm) LLOG1D and LLOG1V FP density (no./ha) and volume (m3) of stage-1 logs (mid-diam. >30 cm) LLOG2D and LLOG2V FP density (no./ha) and volume (m3) of stage-2 logs (mid-diam. >30 cm) LLOG3D and LLOG3V FP density (no./ha) and volume (m3) of stage-3 logs (mid-diam. >30 cm) LLOG4D and LLOG4V FP density (no./ha) and volume (m3) of stage-4 logs (mid-diam. >30 cm) LOG5TL FP total length (m) of stage-5 logs SHRBD FP density (no. /ha) of shrubs (species specific) >0.5 m tall CSAPLD FP density (no. /ha) of conifer saplings. DSAPLD FP density (no. /ha) of deciduous saplings. STUMPD FP density (no. /ha) of tree stumps. LC20BA VR basal area (m2 /ha) of live conifer trees using 20 BAF (species specific) LH20BA VR basal area (m2 /ha) of live deciduous trees using 20 BAF (species specific) DC20BA VR basal area (m2 /ha) of dead conifer trees using 20 BAF (species specific) DH20BA VR basal area (m2 /ha) of dead deciduous trees using 20 BAF (species specific) LC20ST VR no. of stems of live conifer trees using 20 BAF (species; size specific) LH20ST VR no. of stems of live deciduous trees using 20 BAF (species; size specific) DC20ST VR no. of stems of dead conifer trees using 20 BAF (species; size specific) DH20ST VR no. of stems of dead deciduous trees using 20 BAF (species; size specific) a PT = point; PI = point-intercept; FP = fixed plot (78.5 m2); VR = variable-radius plot.
Table 2.3. Linear m
odels established a priori (original) and after initial analysis (exploratory) for predicting abundance (g/ha) of deer m
ice (DM
) in the Sacramento M
ountains, New
Mexico during sum
mers of 1992–1994. R
egressor acronyms are defined in the text.
Model Set /
Acronym
Model Structure (sum
mer biom
ass of deer mice =)
Implied Effects
Original /
DM
Global
$0 + $
1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CG
FHT + $
4CFW
TEMP +
Food supply, fall-winter tem
peratures, availability of $
5CM
ATFEM
+ $6CC
OV
ER !
$7CIN
FCO
MP
mature fem
ales, cover from predation and the
thermal environm
ent, and interspecific interferencecom
petition influence summ
er biomass of deer m
ice..D
MC
1$
0 + $1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CG
FHT + $
4CFW
TEMP !
$5CIN
FCO
MP
Food supply and fal-winter tem
peratures, andinterference com
petition.D
MC
2$
0 + $1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CG
FHT + $
4CFW
TEMP !
$5CD
END
EPFood supply, fall-w
inter temperatures, and delayed
density dependence.D
MC
3$
0 + $1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CG
FHT + $
4CFW
TEMP + $
5CM
ATFEM
Food supply, fall-winter energy dem
and, andavailability of m
ature females.
DM
C4
$0 + $
1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CG
FHT + $
4CC
OV
ER + $
5CA
LTPREY
Food supply, cover, and release from predation.
DM
C5
$0 + $
1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CG
FHT !
$4CD
END
EP + $5CA
LTPREY
Food supply, delayed density dependence, andrelease from
predation.D
MC
6$
0 + $1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CM
ATFEM
+ $4CC
OV
ERFood supply, availability of m
ature females, and
cover.D
MC
7$
0 + $1CM
CO
NIF + $
2CG
FHT + $
3CC
OV
ERM
icrohabitat model.
DM
C8
$0 + $
1CM
CO
NIF + $
2CPR
EVR
AIN
+ $3CFW
TEMP
Food supply and fall-winter tem
peratures.
Table 2.3. (C
ontinued).
Model Set /
Acronym
Model Structure (sum
mer biom
ass of deer mice =)
Implied Effects
Exploratory /
DM
E1$
0 + $1CPR
EVR
AIN
+ $2CG
FHT + $
3CFW
TEMP
Reduced m
odel of food supply and fall-winter
temperatures.
DM
E2$
0 + $1CG
FHT + $
2CFW
TEMP + $
3CM
ATFEM
Reduced m
odel of food supply, fall-winter
temperatures and availability of m
ature females.
DM
E3$
0 + $1CPR
EVR
AIN
+ $2CFW
TEMP + $
3CM
ATFEM
Reduced m
odel of food supply, fall-winter
temperatures and availability of m
ature females.
DM
E4$
0 + $1CPR
EVR
AIN
+ $2CFW
TEMP
Reduced m
odel of food supply and fall-winter
temperatures.
DM
E5$
0 + $1CFW
TEMP
Fall-winter tem
peratures.D
ME6
$0 + $
1CR
SPRA
IN + $
2CFW
TEMP
Food and water supply (indirect) and fall-w
intertem
peratures.D
ME7
$0 + $
1CM
ATFEM
Availability of m
ature females.
Table 2.4. Linear m
odels established a priori (original) and after initial analysis (exploratory) for predicting abundance (g/ha) ofbrush m
ice (BM
) in the Sacramento M
ountains, New
Mexico during sum
mers of 1992–1994. R
egressor acronyms are defined in the
text.
Model Set /
Acronym
Model Structure (sum
mer biom
ass of brush mice =)
Implied Effects
Original /
BM
Global
$0 + $
1CPJTR
EE + $2CPR
EVR
AIN
+ $3CSH
RU
BS + $
4CFW
TEMP +
Food supply, fall-winter tem
peratures, availability of $
5CM
ATFEM
+ $6CR
OC
KS !
$7CIN
FCO
MP
mature fem
ales, cover from predation and therm
alenvironm
ent, and interspecific interferencecom
petition for den sites influence summ
er biomass
of brush mice.
BM
C1
$0 + $
1CPJTR
EE + $2CPR
EVR
AIN
+ $3CSH
RU
BS + $
4CR
OC
KS +
Fooda and presence of suitable den sites as m
odified $
5CLSC
OV!
$6CIN
FCO
MP
by interference competition.
BM
C2
$0 + $
1CO
AK
S + $2CPR
EVR
AIN
+ $3CG
FHT + $
4CR
OC
KS +
Foodb and presence of suitable den sites as m
odified $
5CLSC
OV
! $
6CIN
FCO
MP
by competition.
BM
C3
$0 + $
1CPJTR
EE + $2CPR
EVR
AIN
+ $3CSH
RU
BS + $
4CR
SPRA
IN + $
5CFW
TEMP
Fooda supply, w
ater, and fall-winter tem
peratures.B
MC
4$
0 + $1CO
AK
S + $2CPR
EVR
AIN
+ $3CG
FHT + $
4CFW
TEMP
Foodb supply and fall-w
inter temperatures..
BM
C5
$0 + $
1CPJTR
EE + $2CPR
EVR
AIN
+ $3CSH
RU
BS + $
4CR
SPRA
IN +
Fooda, w
ater, cover and den sites. $
5CR
OC
KS + $
6CLSC
OV
BM
C6
$0 + $
1CO
AK
S + $2CPR
EVR
AIN
+ $3CG
FHT + $
4CR
OC
KS + $
5CLSC
OV
Foodb supply, cover, and den sites.
BM
C7
$0 + $
1CPJTR
EE + $2CPR
EVR
AIN
+ $3CSH
RU
BS + $
4CR
SPRA
IN +
Fooda supply, w
ater, cover, and release from $
5CR
OC
KS + $
6CA
LTPREY
predation.
BM
C8
$0 + $
1CO
AK
S + $2CPR
EVR
AIN
+ $3CG
FHT + $
4CR
OC
KS +
Foodb supply, cover, den sites and release from
$5CLSC
OV
+ $6CA
LTPREY
predation.
BM
C9
$0 + $
1CO
AK
S + $2CPR
EVR
AIN
+ $3CG
FHT + $
4CR
OC
KS + $
5CA
LTPREY
Foodb supply, cover, and release from
predation.B
MC
10$
0 + $1CPJTR
EE + $2CPR
EVR
AIN
+ $3CR
SPRA
IN + $
4CM
ATFEM
+Food from
conifers, water, availability of m
ature $
5CR
OC
KS + $
6CLSC
OV
fem
ales, cover and dens.
Table 2.4. (C
ontinued).
Model Set /
Acronym
Model Structure (sum
mer biom
ass of brush mice =)
Implied Effects
Original /
BM
C11
$0 + $
1CO
AK
S + $2CPR
EVR
AIN
+ $3CG
FHT + $
4CM
ATFEM
+Food
b, water, availability of m
ature females, cover,
$5CR
OC
KS + $
6CLSC
OV
and dens. B
MC
12$
0 + $1CPJTR
EE + $2CSH
RU
BS + $
3CR
OC
KS + $
4CLSC
OV
Microhabitat.
BM
C13
$0 + $
1CSH
RU
BS + $
2CM
ATFEM
+ $3CR
OC
KS
Food from shrubs, availability of fem
ales, and cover.
Exploratory /B
ME1
$0 + $
1CPR
EVR
AIN
+ $2CSH
RU
BS + $
3CM
ATFEM
+ $4CR
OC
KS
Food (shrub mast), availability of m
ature females,
and cover.B
ME2
$0 + $
1CPR
EVR
AIN
+ $2CO
AK
S + $3CM
ATFEM
+ $4CR
OC
KS
Food (acorns), availability of mature fem
ales, andcover.
BM
E3$
0 + $1CM
ATFEM
+ $2CR
OC
KS + $
3CIN
FCO
MP
Availability of m
ature females, cover, and
interference competition for rock dens.
BM
E4$
0 + $1CSH
RU
BS + $
2CM
ATFEM
+ $3CR
OC
KS
Food (shrub mast), availability of m
ature females,
and cover.B
ME5
$0 + $
1CO
AK
S + $2CR
OC
KS
Food (acorns) and cover; reduced microhabitat
model.
BM
E6$
0 + $1CM
ATFEM
+ $2CR
OC
KS
Availability of m
ature females and cover.
BM
E7$
0 + $1CM
ATFEM
Availability of m
ature females.
BM
E8$
0 + $1CR
OC
KS
Rock cover.
a Food from
conifer seeds, juniper berries, shrub mast, including acorns from
low-grow
ing oaks.b Food as acorns from
low-grow
ing and tree oaks, and from grass or forb seeds and foliage.
Table 2.5. Linear m
odels established a priori (original) and after initial analysis (exploratory) for predicting abundance (g/ha) ofM
exican voles (MV
) in the Sacramento M
ountains, New
Mexico during sum
mers of 1992–1994. R
egressor acronyms are defined in
the text.
Model Set /
Acronym
Model Structure (sum
mer biom
ass of Mexican voles =)
Implied Effects
Original /
MV
Global
$0 + $
1CFW
TEMP + $
2CG
FHT + $
3CR
SPRA
IN + $
4CM
ATFEM
+ $5CIN
FCO
MP +
Fall-winter tem
peratures, summ
er cover, food, and $
6CA
LTPREY
! $
7CD
ENSD
EP w
ater supply, availability of sexually mature fem
ales,interspecific interference com
petition for meadow
sites, release from predation, and delayed density-
dependence influences summ
er abundance ofM
exican voles.M
VC
1$
0 + $1CFW
TEMP + $
2CG
FHT + $
3CR
SPRA
IN + $
4CM
ATFEM
+ $5CA
LTPREY
!Fall-w
inter temperatures, cover, food, and w
ater $
6CD
ENSD
EP supply, availability of m
ature females, release from
predation and delayed density-dependence.M
VC
2$
0 + $1CG
FHT + $
2CR
SPRA
IN + $
3CM
ATFEM
! $
4CD
ENSD
EPSum
mer cover, food, and w
ater supply, availabilityof m
ature females, and delayed density-dependence.
MV
C3
$0 + $
1CFW
TEMP + $
2CG
FHT + $
3CR
SPRA
IN + $
4CA
LTPREY
Fall-winter tem
peratures, summ
er cover, food andw
ater supply, and release from predation.
MV
C4
$0 + $
1CFW
TEMP + $
2CG
FHT + $
3CR
SPRA
INFall-w
inter temperature effects on energy dem
and,sum
mer cover, food and w
ater supply.M
VC
5$
0 + $1CG
FHT + $
2CR
SPRA
IN + $
3CM
ATFEM
+ $4CIN
FCO
MP
Summ
er cover, food, and water supply, availability
of mature fem
ales, effects of interferencecom
petition.M
VC
6$
0 + $1CG
FHT + $
2CR
SPRA
IN + $
3CM
ATFEM
Summ
er cover, food, water supply, and availability
of mature fem
ales.
Table 2.5. (C
ontinued).
Model Set /
Acronym
Model Structure (sum
mer biom
ass of Mexican voles =)
Implied Effects
Original /
MV
C7
$0 + $
1CG
FHT + $
2CM
ATFEM
+ $3CIN
FCO
MP
Summ
er cover and food supply, availability ofm
ature females, and effects of interference
competition.
MV
C8
$0 + $
1CFW
TEMP + $
2CG
FHT + $
3CA
LTPREY
Fall-winter tem
peratures, summ
er food and cover,and release from
predation.M
VC
9$
0 + $1CPSSR
AIN
+ $2CM
ATFEM
+ $3CIN
FCO
MP
Previous summ
er-spring rain as an indirect effect onfood and cover, availability of m
ature females, and
effects of interference competition.
MV
C10
$0 + $
1CG
FHT
Summ
er food and cover; microhabitat effects m
odel.
Exploratory /M
VE1
$0 + $
1CG
FHT + $
2CM
ATFEM
Summ
er food, cover, and availability of mature
females.
MV
E2$
0 + $1CFW
TEMP + $
2CG
FHT + $
3CM
ATFEM
+ $4CIN
FCO
MP
Fall-winter tem
peratures, food, cover, availability ofm
ature females, and interference com
petition.M
VE3
$0 + $
1CG
FHT + $
2CM
ATFEM
! $
3CD
ENSD
EPSum
mer food, cover, availability of m
ature females,
and delayed density dependence.M
VE4
$0 + $
1CG
FHT-Q
+ $2CM
ATFEM
-Q + $
3CD
ENSD
EP-QSum
mer food, cover, availability of fem
ales anddelayed density dependence (m
odeled as quadraticfunction).
MV
E5$
0 + $1CM
ATFEM
Availability of m
ature females.
Table 2.6. Linear m
odels established a priori (original) and after initial analysis (exploratory) for predicting abundance (g/ha) oflong-tailed voles (LV
) in the Sacramento M
ountains, New
Mexico during sum
mers of 1992–1994. R
egressor acronyms are defined in
the text.
Model Set /
Acronym
Model Structure (sum
mer biom
ass of long-tailed voles =)Im
plied Effects
Original /
LVG
lobal$
0 + $1CFW
TEMP + $
2CSH
RU
BS + $
3CG
FHT + $
4CM
ATFEM
+ $5CLO
GS +
Fall-winter tem
peratures, food supply, cover, $
6CIN
FCO
MP + $
7CA
LTPREY
availability of sexually m
ature females, interspecific
interference competition and release from
predationinfluences sum
mer biom
ass of long-tailed voles.LV
C1
$0 + $
1CFW
TEMP + $
2CPSSR
AIN
+ $3CSH
RU
BS + $
4CG
FHT + $
5CIN
FCO
MP
Fall-winter tem
peratures, previous summ
er-springrainfall, food supply, cover, and interferencecom
petition.LV
C2
$0 + $
1CFW
TEMP + $
2CSH
RU
BS + $
3CG
FHT + $
4CM
ATFEM
+ $5CIN
FCO
MP
Fall-winter tem
peratures, food supply, cover,availability of m
ature females, and interference
competition.
LVC
3$
0 + $1CSH
RU
BS + $
2CG
FHT + $
3CM
ATFEM
! $
4CD
END
EPFood supply, cover, availability of m
ature females,
and delayed density-dependence. LV
C4
$0 + $
1CSH
RU
BS-Q
+ $2CG
FHT-Q
+ $3CM
ATFEM
-Q !
$4CD
END
EP-QFood supply, cover, availability of m
ature females,
and delayed density-dependence (quadratic function). LV
C5
$0 + $
1CSH
RU
BS + $
2CG
FHT + $
3CM
ATFEM
+ $4CLO
GS + $
5CA
LTPREY
Food supply, cover including logs, availability ofm
ature females, and release from
predation.LV
C6
$0 + $
1CFW
TEMP + $
2CG
FHT + $
3CM
ATFEM
+ $4CLO
GS
Fall-winter tem
peratures, forb-grass supply,availability of m
ature females, and log cover.
LVC
7$
0 + $1CPSSR
AIN
+ $2CSH
RU
BS + $
3CLO
GS
Previous summ
er-spring rainfall, food supply fromshrubs, and log cover.
Table 2.6. (C
ontinued).
Model Set /
Acronym
Model Structure (sum
mer biom
ass of long-tailed voles =)Im
plied Effects
Original /
LVC
8$
0 + $1CSH
RU
BS + $
2CG
FHT + $
3CLO
GS + $
4CA
LTPREY
Food supply, cover, and release from predation.
LVC
9$
0 + $1CSH
RU
BS-Q
+ $2CG
FHT-Q
+ $3CLO
GS-Q
Food supply, and log cover; microhabitat effects
model (quadratic function).
LVC
10$
0 + $1CSH
RU
BS-Q
+ $2CR
SPRA
IN-Q
+ $3CLO
GS-Q
Food from shrubs, extra w
ater from recent spring
rain, and log cover (quadratic function).Exploratory /
LVE1
$0 + $
1CFW
TEMP + $
2CG
FHT + $
3CM
ATFEM
Fall-winter tem
peratures, food and cover from forbs
and grasses, and availability of mature fem
ales.LV
E2$
0 + $1CFW
TEMP + $
2CG
FHT + $
3CM
ATFEM
+ $4CIN
FCO
MP
Fall-winter tem
peratures, food, cover, availability ofm
ature females, and interference com
petition.LV
E3$
0 + $1CFW
TEMP + $
2CM
ATFEM
+ $3CIN
FCO
MP
Fall-winter tem
peratures, availability of mature
females, and interference com
petition.LV
E4$
0 + $1CG
FHT + $
2CM
ATFEM
+ $3CIN
FCO
MP
Food and cover from forbs and grasses, availability
of females, and interference com
petition.LV
E5$
0 + $1CM
ATFEM
+ $2CIN
FCO
MP
Availability of m
ature females and interference
competition.
LVE6
$0 + $
1CG
FHT + $
2CM
ATFEM
Food and cover from forbs and grasses, and
availability of mature fem
ales.LV
E7$
0 + $1CM
ATFEM
Availability of m
ature females.
Table 2.7. Linear m
odels established a priori (original) and after initial analysis (exploratory) for predicting abundance (g/ha) ofM
exican woodrats (M
W) in the Sacram
ento Mountains, N
ew M
exico during summ
ers of 1992–1994. Regressor acronym
s are definedin the text.
Model Set /
Acronym
Model Structure (sum
mer biom
ass of Mexican w
oodrats =)Im
plied Effects
Original /
MW
Global
$0 + $
1CC
SAPL + $
2CSH
RU
BS !
$3CG
FHT + $
4CM
ATFEM
+Food supply, availability of sexually m
ature females,
$5CR
OC
KS + $
6CLG
LOG
! $
7CD
END
EP cover, and delayed density-dependence influencesum
mer biom
ass of Mexican w
oodrats..M
WC
1$
0 + $1CSH
RU
BS !
$2CG
FHT + $
3CM
ATFEM
+ $4CR
OC
KS !
$5CD
END
EPFood from
shrubs and forbs, availability of mature
females, cover, and delayed density-dependence.
MW
C2
$0 + $
1CM
TMG
OK
! $
2CG
FHT + $
3CM
ATFEM
+ $4CR
OC
KS !
$5CD
END
EPFood from
mountain m
ahogony, Gam
bel oak, andforbs, availability of m
ature females, cover, and
delayed density-dependence.M
WC
3$
0 + $1CC
SAPL + $
2CSH
RU
BS + $
3CM
ATFEM
+ $4CR
OC
KS + $
5CA
LTPREY
Food from conifer C
SAPLings and shrubs,
availability of mature fem
ales, cover, and releasefrom
predation. M
WC
4$
0 + $1CSH
RU
BS + $
2CM
ATFEM
+ $3CR
OC
KS + $
4CLG
LOG
+ $5CA
LTPREY
Food from shrubs, availability of m
ature females,
cover, and release from predation.
MW
C5
$0 + $
1CSH
RU
BS + $
2CM
ATFEM
+ $3CR
OC
KS + $
4CLG
LOG
Food from shrubs, availability of m
ature females, and
cover.M
WC
6$
0 + $1CC
SAPL + $
2CSH
RU
BS + $
3CR
SPRA
IN + $
4CR
OC
KS
Food from conifer C
SAPLings and shrubs, w
aterfrom
spring rains, and cover.M
WC
7$
0 + $1CC
SAPL + $
2CM
TMG
OK
+ $3CR
SPRA
IN + $
4CR
OC
KS
Food from conifer C
SAPLings and m
ountainm
ahogany and Gam
bel oak, water from
spring rains,and cover.
Table 2.7. (C
ontinued).
Model Set /
Acronym
Model Structure (sum
mer biom
ass of Mexican w
oodrats =)Im
plied Effects
Original /
MW
C8
$0 + $
1CFW
TEMP + $
2CC
SAPL + $
3CSH
RU
BS !
$4CG
FHT + $
5CR
OC
KS
Fall-winter tem
perature, food supply, and cover.M
WC
9$
0 + $1CFW
TEMP + $
2CSH
RU
BS + $
3CR
SPRA
IN + $
4CR
OC
KS + $
5CLG
LOG
Fall-winter tem
perature, food from shrubs, w
aterfrom
spring rains, and cover.M
WC
10$
0 + $1CM
TMG
OK
+ $2CM
ATFEM
+ $3CR
OC
KS + $
4CLG
LOG
Food from m
ountain mahogany and G
ambel oak,
availability of mature fem
ales, and cover.M
WC
11$
0 + $1CC
SAPL + $
2CSH
RU
BS !
$3CG
FHT + $
4CR
OC
KS + $
5CLG
LOG
Food supply and cover; microhabitat effects m
odel.
Exploratory /M
WE1
$0 + $
1CM
ATFEM
+ $2CR
OC
KS + $
3CLG
LOG
Availability of m
ature females, rock cover, and large
log cover.M
WE2
$0 + $
1CM
ATFEM
+ $2CR
SPRA
IN + $
3CR
OC
KS + $
4CLG
LOG
Availability of m
ature females, m
oisture from recent
spring rains, rock cover and large log cover.M
WE3
$0 + $
1CM
ATFEM
+ $2CR
SPRA
IN + $
3CLG
LOG
Availability of m
ature females, m
oisture from recent
spring rains, and large log cover.M
WE4
$0 + $
1CM
ATFEM
+ $2CR
OC
KS
Availability of fem
ales, and rock cover.M
WE5
$0 + $
1CM
ATFEM
+ $2CLG
LOG
Availability of m
ature females and large log cover
MW
E6$
0 + $1CM
ATFEM
+ $2CR
SPRA
INA
vailability of mature fem
ales and moisture from
recent spring rains.M
WE7
$0 + $
1CM
ATFEM
Availability of m
ature females.
186
Table 2.8. Mean summer density ( /ha), mass (g), and biomass ( ; g/ha)
of five murid rodents occurring in three habitats (mesic forest, montane meadow, and xeric forest) of the Sacramento Mountains, New Mexico (1991–1996).
b c
Species (i) Habitat (k) a ( /ha) (g) (g/ha) (g/ha) (%) n1, n2, d
Deer mouse Mesic Forest-L 7.8 15.2 125.5 10.79 9 5, 5, 39Mesic Forest-M 10.6 17.4 175.7 8.78 5 23, 23, 45Montane Meadow 8.4 18.8 157.5 13.45 9 22, 21, 20Xeric Forest 4.9 17.4 84.6 6.18 7 25, 24, 25
Brush mouse Mesic Forest-L 0.1 20.0 1.0 — — 5, 0, 1
Mesic Forest-M 0.3 23.8 8.1 13.21 209 23, 3, 2Xeric Forest 5.8 23.6 138.6 7.22 5 25, 21, 29
Mexican vole Mesic Forest-L 0.4 27.5 10.3 15.33 149 5, 1, 2Mesic Forest-M 1.4 25.6 38.8 39.36 101 23, 4, 4Montane Meadow 59.8 28.4 1739.7 71.61 4 22, 20, 85Xeric Forest 1.5 23.3 38.0 9.65 25 25, 7, 7
Long-tailed Mesic Forest-L 3.3 31.2 104.1 14.32 14 5, 4, 15vole Mesic Forest-M 6.4 33.3 213.5 38.01 18 23, 15, 20
Montane Meadow 11.2 34.9 373.0 70.23 19 22, 17, 18Xeric Forest < 0.1 33.0 0.7 — — 25, 0, 1
Mexican Mesic Forest-L 0.7 122.4 81.2 28.36 35 5, 3, 5woodrat Mesic Forest-M 0.9 120.2 106.1 17.12 16 23, 16, 6
Montane Meadow 0.2 113.9 28.9 53.88 187 22, 2, 2Xeric Forest 1.1 119.6 132.7 14.09 11 25, 21, 7
a For Mesic Forest, L = late seral stage (dominant conifers > 200 yrs old); M = mid-seral stage (dominant conifers 60–100 yrs old).b The square root of the average enumeration error of within each habitat or habitat seral stage.
c CVe is the percent of variation in attributed to enumeration error, relative to the magnitude of .
d n1 is the number of replicates (sites and years) used to estimate mean density and mean biomass; n2 is the number of
replicates where enough captures permitted estimation of the sampling variance of biomass; is the mean number of
individuals captured and marked per replicate. Mean values of density and include zero values for sites or years
where a species was not captured except for brush mice which were never captured in montane meadows and wereexcluded.
Table 2.9. Tem
poral and spatial effects on summ
er biomass (g/ha) of five m
urid rodents occurring in three habitats (m
id-seral mesic forest, m
ontane meadow
, and xeric forest) of the Sacramento M
ountains, New
Mexico (1992–1995).
Long-tailed M
exican D
eer mouse
Brush m
ouse M
exican vole vole
woodrat
Effect ad.f. b
F P
F P
F P
F P
F P
Habitat
2, 7 1.00
0.414 7.29
0.019 4.40
0.058 2.48
0.15428.04
0.001
Year
3, 7 9.74
0.00721.57
0.007 7.91
0.012 5.14
0.034 4.01
0.059
Habitat x Y
ear6, 7
1.750.241
17.380.007
7.500.009
4.270.039
0.940.522
a The effects w
ere quantified as F-ratios between variances estim
ated with a m
ixed model of repeated m
easures of biomass over tim
e and assuming
unstructured covariances.b N
umerator and denom
inator degrees of freedom.
Table 2.10. M
ean biomass (g/ha) of five m
urid rodents occurring in mid-seral m
esic forests and xeric forests of the Sacram
ento Mountains, N
ew M
exico during July–August 1991 and 1992.
n1 b
n2
Species / habitat1991
19921991
19921991
19921991
1992 Z
c P
Deer m
ousem
id-seral mesic forest
783 85
47.1 10.6
2 6
169 23
14.44<0.001
xeric forest 89
110 33.4
8.3 2
6 24
33 0.61
0.541
Brush m
ousem
id-seral mesic forest
58 1
19.4 31.5
1 1
12 1
1.56 0.119
xeric forest 44
240 14.6
16.5 1
5 10
45 8.93
<0.001
Mexican vole
mid-seral m
esic forest 95
104 66.4
47.6 1
3 5
7 0.11
0.915xeric forest
74 121
30.2 13.6
1 4
8 19
1.41 0.158
Long-tailed volem
id-seral mesic forest
432 516
191.2103.4
2 4
16 49
0.39 0.700
Mexican w
oodratm
id-seral mesic forest
55 146
97.8 40.3
1 5
2 5
0.86 0.387
xeric forest 69
185 14.3
28.0 1
6 4
9 3.67
<0.001 a Square root of the enum
eration variance averaged across site replicates.b n
1 = number of sites w
here enough animals w
ere captured and marked to estim
ate enumeration error; n
2 = mean num
ber of marked anim
als used to estim
ate biomass at a given site.
c Z tests Ho : no difference in biom
ass between 1991 and 1992 w
ithin a habitat.
Table 2.11. B
iomass (g/ha) of five m
urid rodents occurring at a mid- and one late-seral m
esic forest site, one montane
meadow
site, and one xeric forest site in the Sacramento M
ountains, New
Mexico during July–A
ugust 1995 and 1996.
n a
Species / seral stage1995
19961995
19961995
1996 Z
1 b P
1 Z
2 c P
2
Deer m
ouselate-seral m
esic forest 52
182 13.8
27.5 22
524.24
<0.0011.72
0.086m
id-seral mesic forest
96 260
15.4 36.0
34 72
4.19<0.001
montane m
eadow 204
305 48.1
55.3 19
421.38
0.167xeric forest
53 183
9.9 37.0
22 49
3.41 0.001
Brush m
ousexeric forest
11 71
2.9 16.5
3 19
3.56<0.001
Mexican vole
montane m
eadow4184
2331338.1
353.1244
1073.79
<0.001
Long-tailed volelate-seral m
esic forest 36
390 17.2
51.6 5
526.53
<0.0012.19
0.028m
id-seral mesic forest
48 241
15.4 44.3
7 27
4.12<0.001
Mexican w
oodratlate-seral m
esic forest 80
239 58.2
47.5 4
112.13
0.0341.70
0.088m
id-seral mesic forest
84 113
57.0 57.0
3 5
0.35 0.725
xeric forest 87
212 33.9
47.2 5
112.15
0.032 a N
umber of m
arked animals used to estim
ate biomass.
b Z1 tests H
o : no difference in biomass betw
een 1995 and 1996 within a seral stage or habitat.
c Z2 tests H
o : no difference in biomass betw
een mesic forest seral stages in 1996.
190
Table 2.12. Seral-stage and year effects on biomass (g/ha) of three rodents occurringin mid- (60–100 yr) and late (>200 yr) seral stages of mesic forest of the SacramentoMountains, New Mexico during July–August 1994 and 1995.
Deer Mouse Long-tailed vole Mexican woodratEffect a df b F P F P F P
Seral stage 1, 3 1.99 0.253 0.20 0.688 1.78 0.274
Year 1, 3 27.59 0.013 * 0.12 0.752 0.53 0.514
Stage x Year 1, 3 0.65 0.478 <0.01 0.952 0.84 0.428 a Effects were tested using a repeated measures, mixed model assuming unstructured covariances.b Numerator, denominator (mid seral = 3 sites; late seral = 2 sites).* Significant at the " = 0.05 level.
Table 2.13. C
oefficients of modeled effects on sum
mer abundance (g/ha) of deer m
ice in the Sacramento M
ountains, New
Mexico
(1992–1994) and weights of m
odel reliability. Sample size (n = 42) is based on 20 sites each sam
pled during 1, 2 or 3 summ
ers. M
odel order based on )A
ICc . R
egressor variables are described in the text and Table 2.3.
DM
aR
2 O
riginal Models
c A
ll Models
d PR
EV-
MA
T-M
odelK
b(%
))
AIC
cW
eight)
AIC
c W
eightIN
TERC
EPTM
CO
NIF
RA
ING
FHT
FWTEM
PFEM
CO
VER
E25
53.5–
–0
0.6701*!
127.768–
–2.226
!26.267
8.214–
E35
49.1–
–3.73
0.1037*!
55.668–
!1.779
–!
39.2197.272
–G
lobal9
61.50
0.4822*3.95
0.0931*22.258
!0.146
!2.509
!2.948
!35.900
8.110!
0.0012C
37
55.00.20
0.4355*4.15
0.0841*!
121.657!
0.500!
0.5491.584
!34.625
8.188–
E73
39.2–
–6.17
0.0306 50.617
––
––
9.479–
C6
647.8
3.590.0800*
7.540.0155
!195.406
!0.598
3.824–
–9.591
!0.0029
E53
29.3–
–12.51
0.0013!
218.138–
––
!52.350
––
E64
30.3–
–14.34
0.0005!
201.316–
––
!45.742
––
E44
30.1–
–14.46
0.0005!
90.604–
!2.568
–!
61.323–
–C
85
32.111.89
0.001315.84
0.0002!
181.098!
0.556!
1.171–
!62.167
––
E15
31.3–
–16.31
0.0002!
82.965–
!2.656
1.100 !
59.540–
–C
17
39.812.45
0.001016.39
0.0002!
46.978!
0.138!
3.145!
4.206!
63.095–
–C
27
33.716.50
0.000120.45
<0.0001!
209.643!
0.408!
0.6840.602
!60.822
––
C7
5 7.1
25.04<0.0001
28.98<0.0001
91.4010.506
–2.052
––
!0.0022
C5
716.3
26.31<0.0001
30.26<0.0001
!472.962
!0.374
8.3082.183
––
–C
47
14.527.19
<0.000131.14
<0.0001!
404.523!
0.3977.292
2.003–
–!
0.0004
Table 2.13. (C
ontinued).
DM
aR
2 O
riginal Models c
All M
odelsd
Model
K b
(%)
)A
ICc
Weight
)A
ICc
Weight
INFC
OM
PA
LTPREY
DEN
DEP
SPRR
AIN
E25
53.5–
–0
0.6701*–
––
–E3
549.1
––
3.730.1037*
––
––
Global
961.5
00.4822*
3.950.0931*
0.879–
––
C3
755.0
0.200.4355*
4.150.0841*
––
––
E73
39.2–
–6.17
0.0306–
––
–C
66
47.83.59
0.0800*7.54
0.0155–
––
–E5
329.3
––
12.510.0013
––
––
E64
30.3–
–14.34
0.0005–
––
2.244E4
430.1
––
14.460.0005
––
––
C8
532.1
11.890.0013
15.840.0002
––
––
E15
31.3–
–16.31
0.0002–
––
–C
17
39.812.45
0.001016.39
0.00020.955
––
–C
27
33.716.50
0.000120.45
<0.0001–
–!
0.848–
C7
5 7.1
25.04<0.0001
28.98<0.0001
––
––
C5
716.3
26.31<0.0001
30.26<0.0001
–!
0.009!
0.949–
C4
714.5
27.19<0.0001
31.14<0.0001
–!
0.009–
– a Linear regression m
odels of deer mouse abundance described in Table 2.3.
b Num
ber of model param
eters estimated from
data including intercept coefficient, effect coefficients, and a residual error term.
c Weighted according to m
inimum
AIC
c value of the nine a priori models.
d Weighted according to m
inimum
AIC
c value of the nine a priori and seven exploratory models.
193
Table 2.14. Precision and relative importance of ecological factors associated withsummer abundance (g/ha) of deer mice in the Sacramento Mountains, New Mexico(1992–1994). Two sets of regression coefficients ( ) are given; those averaged over apriori (original) models and the other averaged over original and exploratory (all) models. CVs of the average coefficients include variation associated with uncertainty ofmodel structure. Relative importance (RI) of each regressor variable is also shown forboth model sets and is based on the sums of Akaike weights for each model that includedthe regressor. Asterisks indicate effects included in the 95% confidence set of models. Regressor acronyms are defined in the text. Factors are ordered from highest to lowestaccording to RI in the original model set.
Original Models All Models Factor CV(%) RI CV(%) RI
MCONIF !0.337 164.1 1.000* !0.065 181.4 0.193*PREVRAIN !1.147 329.7 1.000* !0.408 274.1 0.297*MATFEM 8.243 23.3 0.998* 8.141 23.7 0.997*FWTEMP !32.535 47.5 0.920* !28.083 48.7 0.954*GFHT !0.736 366.9 0.919* 1.349 113.2 0.848*COVER !0.0008 189.2 0.562* !0.0002 203.8 0.109*INFCOMP 0.425 66.7 0.483* 0.082 100.0 0.093*DENDEP !0.0001 151.1 0.000 !0.00002 151.1 0.000ALTPREY .0.a 163.2 0.000 .0.a 163.2 0.000RSPRAIN – – – 0.001 165.4 0.001 a #0.000001
Table 2.15. C
oefficients of modeled effects on sum
mer abundance (g/ha) of brush m
ice in the Sacramento M
ountains, New
Mexico
(1992–1994) and weights of m
odel reliability. Sample size (n = 42) is based on 20 sites each sam
pled during 1, 2 or 3 summ
ers. M
odel order based on )A
ICc . R
egressor variables are described in the text and Table 2.4.
BM
aR
2O
riginal Models
c A
ll Models
d IN
TER-
PREV
-FW
-R
SP-M
odelK
b(%
))
AIC
cW
eight)
AIC
c W
eightC
EPTPJTR
EER
AIN
TEMP
RA
INSH
RU
BS
GFH
T E4
593.5
––
00.3414*
10.159–
––
––
–E7
392.6
––
0.330.2902*
3.942–
––
––
–E6
492.8
––
1.810.1382*
!6.326
––
––
––
E26
93.5–
–2.66
0.0902*!
32.300–
0.572–
––
–E3
593.0
––
2.930.0789*
!9.804
––
––
––
C13
592.8
00.8547*
4.350.0389*
!6.161
––
––
0.0004–
E16
92.9–
–6.17
0.0156134.993
–!
1.842–
–0.0007
–C
118
93.5 4.19
0.1052*8.53
0.0048!
17.350–
0.425–
––
!0.202
C10
893.1
6.950.0265
11.290.0012
156.7360.043
!2.193
–0.430
––
Global
993.4
8.280.0136
12.630.0006
180.8190.038
!2.570
!1.406
–0.0012
–E5
463.4
––
69.95<0.0001
!48.780
––
––
––
E83
58.5–
–72.85
<0.0001!
99.384–
––
––
–C
28
67.971.45
<0.000175.80
<0.0001 100.102
–!
2.352–
––
1.087C
18
67.771.73
<0.000176.07
<0.0001440.910
!0.173
!7.075
––
0.0026–
C4
662.6
71.85<0.0001
76.20<0.0001
37.629–
!1.255
!4.879
––
0.144C
67
63.973.32
<0.000177.67
<0.0001 1.189
–!
1.027–
––
1.190C
97
63.873.41
<0.000177.76
<0.0001!
40.501–
!0.352
––
–0.878
C12
660.5
74.15<0.0001
78.49<0.0001
!91.703
!0.132
––
–0.0024
–C
78
64.875.30
<0.000179.64
<0.0001368.095
!0.126
!7.061
–4.570
0.0019–
C5
864.6
75.52<0.0001
79.87<0.0001
442.653!
0.137!
7.981–
4.6010.0004
–C
88
63.976.37
<0.000180.72
<0.00014.208
–!
1.100–
––
1.084C
37
57.879.82
<0.000184.16
<0.0001999.615
0.077!
11.74035.237
7.5920.0091
–
Table 2.15. (C
ontinued).
BM
aR
2O
riginal Models
c A
ll Models
d M
AT-
INF-
ALT-
Model
K b
(%)
)A
ICc
Weight
)A
ICc
Weight
OA
KS
FEMR
OC
KS
LSCO
VC
OM
PPR
EY E4
593.5
––
00.3414*
0.01116.058
!0.779
––
–E7
392.6
––
0.330.2902*
–17.547
––
––
E64
92.8–
–1.81
0.1382* –
16.6540.432
––
–E2
693.5
––
2.660.0902*
0.01216.076
!0.790
––
–E3
593.0
––
2.930.0789*
–16.313
0.408–
7.112–
C13
592.8
00.8547*
4.350.0389*
–16.646
0.388–
––
E16
92.9–
–6.17
0.0156–
16.4320.077
––
–C
118
93.5 4.19
0.1052*8.53
0.00480.012
16.119!
0.891!
0.0003–
–C
108
93.16.95
0.0265 11.29
0.0012–
16.635!
0.1520.0168
––
Global
993.4
8.280.0136
12.630.0006
–16.197
!0.361
–7.912
–E5
463.4
––
69.95<0.0001
0.027–
1.640–
––
E83
58.5–
–72.85
<0.0001–
–5.165
––
–C
28
67.971.45
<0.000175.80
<0.00010.022
–1.583
0.006028.436
–C
18
67.771.73
<0.000176.07
<0.0001–
–3.854
!0.0574
35.555–
C4
662.6
71.85<0.0001
76.20<0.0001
0.037–
––
––
C6
763.9
73.32<0.0001
77.67<0.0001
0.024–
2.1740.0246
––
C9
763.8
73.41<0.0001
77.76<0.0001
0.025–
2.096–
–0.0005
C12
660.5
74.15<0.0001
78.49<0.0001
––
5.482!
0.0712–
–C
78
64.875.30
<0.000179.64
<0.0001–
–4.800
––
0.0045C
58
64.675.52
<0.000179.87
<0.0001–
–4.751
0.0274–
–C
88
63.976.37
<0.000180.72
<0.00010.024
–2.197
0.0278–
0.0014C
37
57.879.82
<0.000184.16
<0.0001–
––
––
– a Linear regression m
odels of brush mouse abundance described in Table 2.4.
b Num
ber of model param
eters estimated from
data, including intercept coefficient, effect coefficients, and a residual error term.
c Weighted according to m
inimum
AIC
c value of the 14 a priori models.
d Weighted according to m
inimum
AIC
c value of the 14 a priori and eight exploratory models.
196
Table 2.16. Precision and relative importance of ecological factors associated withsummer abundance (g/ha) of brush mice in the Sacramento Mountains, New Mexico(1992–1994). Two sets of regression coefficients ( ) are given; those averaged over apriori (original) models and the other averaged over original and exploratory (all) models. CVs of the average coefficients include variation associated with uncertainty ofmodel structure. Relative importance (RI) of each regressor variable is also shown forboth model sets and is based on the sums of Akaike weights for each model that includedthe regressor. Asterisks indicate effects included in the 95% confidence set of models. Regressor acronyms are defined in the text. Factors are ordered from highest to lowestaccording to RI in the original model set.
Original Models All Models Factor CV(%) RI CV(%) RI
MATFEM 16.584 6.8 1.000* 16.624 7.3 1.000*ROCKS 0.229 354.6 1.000* !0.234 299.6 0.710*SHRUBS 0.0003 11,095.0 0.868* 0.00002 9,020.6 0.055PREVRAIN !0.048 595.4 0.145* 0.020 945.1 0.112*LSCOV 0.0004 6,168.9 0.132* 0.00002 6,169.1 0.006GFHT !0.021 458.8 0.105* !0.001 460.9 0.005OAKS 0.001 699.3 0.105* 0.005 681.8 0.436*PJTREE 0.002 581.9 0.040 0.00008 582.5 0.002RSPRAIN 0.011 283.0 0.026 0.0005 283.9 0.001INFCOMP 0.107 103.9 0.014 0.566 98.4 0.080*FWTEMP !0.019 239.8 0.014 !0.0009 240.3 0.001ALTPREY .0.a 6,420.0 0.000 .0.a 6,420.0 0.000 a #0.000001
Table 2.17. C
oefficients of modeled effects on sum
mer abundance (g/ha) of M
exican voles in the Sacramento M
ountains, New
M
exico (1992–1994) and weights of m
odel reliability. Sample size (n = 42) is based on 20 sites each sam
pled during 1, 2 or 3 sum
mers. M
odel order based on )A
ICc . R
egressor variables are described in the text and Table 2.5.
MV
aR
2O
riginal Models
c A
ll Models
d IN
TER-
FW-
RSP-
MA
T-IN
F-M
odelK
b(%
))
AIC
cW
eight)
AIC
c W
eightC
EPTTEM
PG
FHT
RA
INFEM
CO
MP
Global
995.5
00.9202*
00.9051*
86.52468.514
76.759!
3.07346.708
35.466C
75
93.25.55
0.0573*5.52
0.0563* !
449.116–
70.484–
35.71431.787
C5
693.4
7.430.0224
7.390.0221
!563.997
–65.872
13.62837.300
30.017E2
693.3
––
8.01 0.0162
!246.431
33.08771.222
–35.375
32.443E3
591.2
––
16.450.0002
!423.390
–89.776
–54.277
–C
26
91.318.60
0.000118.57
0.0001 !
543.283–
83.83013.857
54.065–
C1
892.3
19.62<0.0001
19.51<0.0001
!323.223
61.39181.530
15.01560.016
–C
65
89.723.00
<0.0001 22.96
<0.0001!
748.100–
72.47034.854
42.698–
E14
88.7–
–24.23
<0.0001!
455.142–
86.554–
39.126–
C9
588.7
26.92<0.0001
26.88<0.0001
!376.176
––
–65.920
37.638C
103
83.737.42
<0.0001 37.38
<0.0001!
685.259–
149.040–
––
E45
84.3–
–14.83
<0.0001 !
34.137–
––
––
C4
584.0
41.63<0.0001
41.59<0.0001
!627.210
42.357144.435
22.567–
–C
85
83.742.42
<0.0001 42.38
<0.0001!
788.130!
15.134148.665
––
–E5
381.5
––
42.60<0.0001
!17.029
––
–79.436
–C
36
84.044.34
<0.0001 44.30
<0.0001!
632.79643.965
144.30322.744
––
Table 2.17. (C
ontinued).
MV
aR
2 O
riginal Models
c A
ll Models
d A
LT-PSS-
MA
T- M
odelK
b(%
))
AIC
cW
eight)
AIC
c W
eightPR
EYD
END
EPR
AIN
GFH
T_QFEM
_QD
END
_Q
Global
995.5
00.9202*
00.9051*
!0.121
!6.594
––
––
C7
593.2
5.550.0572*
5.550.0563*
––
––
––
C5
693.4
7.430.0224
7.430.0221
––
––
––
E26
93.3–
–8.05
0.0162–
––
––
– E3
591.2
––
16.480.0002
–!
7.635–
––
– C
26
91.318.60
0.000118.60
0.0001–
!6.812
––
––
C1
892.3
19.62<0.0001
19.62<0.0001
0.302!
8.965–
––
– C
65
89.723.00
<0.0001 23.00
<0.0001 –
––
––
– E1
488.7
––
24.27<0.0001
––
––
––
C9
588.7
26.92<0.0001
26.92<0.0001
––
4.848–
––
C10
383.7
37.42<0.0001
37.42<0.0001
––
––
––
E45
84.3–
–40.80
<0.0001–
––
6.8560.839
0.026C
45
84.041.63
<0.0001 41.63
<0.0001–
––
––
– C
85
83.742.42
<0.0001 42.42
<0.00010.017
––
––
– E5
381.5
––
42.64<0.0001
––
––
––
C3
684.0
44.34<0.0001
44.34<0.0001
0.025–
––
––
a Linear regression m
odels of Mexican vole abundance described in Table 2.5.
b Num
ber of model param
eters estimated from
data including intercept coefficient, effect coefficients, and a residual error term.
c Weighted according to m
inimum
AIC
c value of the 11 a priori models.
d Weighted according to m
inimum
AIC
c value of the 11 a priori and five exploratory models.
199
Table 2.18 Precision and relative importance of ecological factors associated withsummer abundance (g/ha) of Mexican voles in the Sacramento Mountains, New Mexico(1992–1994). Two sets of regression coefficients ( ) are given; those averaged over apriori (original) models and the other averaged over original and exploratory (all) models. CVs of the average coefficients include variation associated with uncertainty ofmodel structure. Relative importance (RI) of each regressor variable is also shown forboth model sets and is based on the sums of Akaike weights for each model that includedthe regressor. Asterisks indicate effects included in the 95% confidence set of models. Regressor acronyms are defined in the text. Factors are ordered from highest to lowestaccording to RI in the original model set.
Original Models All Models Factor CV(%) RI CV(%) RI
GFHT 76.156 17.7 1.000* 76.080 17.8 1.000*MATFEM 45.869 18.3 1.000* 45.701 18.5 1.000*INFCOMP 35.128 20.5 1.000* 35.076 20.5 1.000*RSPRAIN !2.520 636.9 0.943* !2.478 636.9 0.927*DENDEP !6.069 34.1 0.920* !5.970 34.4 0.905*FWTEMP 63.048 104.3 0.920* 62.548 105.5 0.921*ALTPREY !0.111 118.3 0.920* !0.110 118.4 0.905*PSSRAIN 0.000006 279.4 0.000 0.000006 279.4 0.000GFHT_Q – – – .0.a 100.8 0.000MATFEM_Q – – – .0.a 103.6 0.000DEND_Q – – – .0.a 127.3 0.000 a #0.000001
Table 2.19. C
oefficients of modeled effects on sum
mer abundance (g/ha) of long-tailed voles in the Sacram
ento Mountains, N
ew
Mexico (1992–1994) and w
eights of model reliability. Sam
ple size (n = 42) is based on 20 sites each sampled during 1, 2 or
3 summ
ers. Model order based on )
AIC
c . Regressor variables are described in the text and Table 2.6.
LVa
R 2
Original M
odelsc
All M
odelsd
INTER
-FW
-M
AT-
INF-
ALT-
Model
K b
(%)
)A
ICc
Weight
)A
ICc
Weight
CEPT
TEMP
SHR
UB
SG
FHT
FEMLO
GS
CO
MP
PREY
E45
79.7–
–0
0.6172*114.612
––
!26.193
35.014–
8.908–
E26
80.3–
–1.52
0.2883* !
98.388!
34.764 –
!26.723
34.093–
8.911–
C2
780.3
00.7190*
4.410.0680*
!93.142
!34.394
!0.0006
!26.842
34.043–
8.914–
Global
982.3
1.880.2810*
6.29 0.0266
!209.982
!58.583
0.0059!
24.39738.155
!0.402
8.593!
0.030E6
461.7
––
24.14<0.0001
!82.765
––
16.45737.878
––
–E1
562.3
––
26.10<0.0001
!293.856
!34.442
–15.939
36.966–
––
C6
664.2
22.22<0.0001
26.64<0.0001
!406.730
!63.980
–12.470
39.301!
0.396–
–C
36
63.123.49
<0.0001 27.91
<0.0001!
105.570–
0.000716.130
32.941–
4.377–
E54
57.1–
– 28.97
<0.0001 !
22.333–
––
35.072–
––
C5
764.2
25.12<0.0001
29.53<0.0001
!37.181
–0.0033
19.39442.188
!0.342
4.305–
E35
58.8–
–29.83
<0.0001 !
194.292!
27.701–
–34.339
––
–E7
348.8
––
33.91<0.0001
!50.938
––
–42.997
–9.096
–C
17
58.331.51
<0.0001 35.92
<0.0001!
858.922!
51.253!
0.0055!
27.171–
––
–C
46
44.840.40
<0.0001 44.81
<0.0001!
65.932–
0.01022.602
33.389–
–!
0.048C
75
34.744.77
<0.000149.18
<0.0001 56.854
–!
0.0121–
–!
0.288–
–C
86
28.951.06
<0.0001 55.47
<0.0001!
633.475–
!0.0177
21.646–
0.347–
!0.003
C10
522.2
52.10<0.0001
56.51<0.0001
!1.541
–!
0.0100–
––
––
C9
517.9
54.38<0.0001
58.79<0.0001
!633.475
–!
0.041419.732
–0.241
––
Table 2.19. (C
ontinued).
LVa
R 2
Original M
odelsc
All M
odelsd
PSS-D
EN-
MA
T-R
SP-M
odelK
b(%
))
AIC
cW
eight)
AIC
c W
eightR
AIN
DEP
GFH
T_QSH
RB
_QFEM
_QLO
GS_Q
RSPR
_QD
END
_QR
AIN
E45
79.7–
–0
0.6172*–
––
––
––
––
E26
80.3–
–1.52
0.2883*–
––
––
––
––
C2
780.3
00.7187*
4.410.0680*
––
––
––
––
–G
lobal9
82.31.88
0.2813*6.29
0.0266–
––
––
––
––
E64
61.7–
–24.14
<0.0001–
––
––
––
––
E15
62.3–
–26.10
<0.0001–
––
––
––
––
C6
664.2
22.22<0.0001
26.64<0.0001
––
––
––
––
–C
36
63.123.49
<0.0001 27.91
<0.0001–
4.502–
––
––
––
E54
57.1–
–28.97
<0.0001–
––
––
––
– 24.878
C5
764.2
25.12<0.0001
29.53<0.0001
––
––
––
––
–E3
558.8
––
29.83<0.0001
9.128–
––
––
––
–E7
348.8
––
33.91<0.0001
––
––
––
––
–C
17
58.331.51
<0.0001 35.92
<0.00017.659
––
––
––
––
C4
644.8
40.40<0.0001
44.81<0.0001
––
––
––
– 0.268
–C
75
34.744.77
<0.000149.18
<0.000113.507
–1.189
!0.000
1.208–
––
–C
86
28.951.06
<0.0001 55.47
<0.0001–
––
!0.000
––
––
–C
105
22.252.10
<0.0001 56.51
<0.0001–
––
––
0.0009 10.696
––
C9
517.9
54.38<0.0001
58.79<0.0001
––
1.156 !
0.000–
0.0011–
––
a Linear regression m
odels of long-tailed vole abundance described in Table 2.6.b N
umber of m
odel parameters estim
ated from data including intercept coefficient, effect coefficients, and a residual error term
.c W
eighted according to minim
um A
ICc value of the 11 a priori m
odels.d W
eighted according to minim
um A
ICc value of the 11 a priori and seven exploratory m
odels.
202
Table 2.20 Precision and relative importance of ecological factors associated withsummer abundance (g/ha) of long-tailed voles in the Sacramento Mountains, NewMexico (1992–1994). Two sets of regression coefficients ( ) are given; those averagedover a priori (original) models and the other averaged over original and exploratory (all) models. CVs of the average coefficients include variation associated with uncertainty ofmodel structure. Relative importance (RI) of each regressor variable is also shown forboth model sets and is based on the sums of Akaike weights for each model that includedthe regressor. Asterisks indicate effects included in the 95% confidence set of models. Regressor acronyms are defined in the text. Factors are ordered from highest to lowestaccording to RI in the original model set.
Original Models All Models Factor CV(%) RI CV(%) RI
MATFEM 35.200 14.9 1.000* 34.767 13.6 1.000*INFCOMP 8.823 17.6 1.000* 8.901 17.2 1.000*GFHT !26.152 32.6 1.000* !26.343 30.8 1.000*FWTEMP !41.208 91.3 1.000* !13.918 112.7 0.383*SHRUBS 0.0012 748.0 1.000* 0.0001 750.6 0.095*LOGS !0.113 93.4 0.281* !0.011 114.1 0.027 ALTPREY !0.0086 119.9 0.281* !0.0008 136.7 0.027MATFEM_Q .0.a 103.5 0.000 .0.a 103.5 0.000GFHT_Q .0.a 104.3 0.000 .0.a 104.3 0.000RSPR_Q .0.a 104.9 0.000 .0.a 104.9 0.000PSSRAIN .0.a 110.9 0.000 0.000002 106.5 0.000DEND_Q .0.a 112.5 0.000 .0.a 112.5 0.000DENDEP 0.00003 131.2 0.000 0.000002 131.2 0.000LOGS_Q .0.a 184.7 0.000 .0.a 184.7 0.000SHRB_Q .0.a 307.4 0.000 .0.a 307.4 0.000RSPRAIN – – – 0.000008 106.4 0.000 a #0.000001
Table 2.21. C
oefficients of modeled effects on sum
mer abundance (g/ha) of M
exican woodrats in the Sacram
ento Mountains,
New
Mexico (1992–1994) and w
eights of model reliability. Sam
ple size (n = 42) is based on 20 sites each sampled during 1, 2
or 3 summ
ers. Model order based on )
AIC
c . Regressor variables are described in the text and Table 2.7.
MW
aR
2O
riginal Models
c A
ll Models
d IN
TER-
MA
T-M
odelK
b(%
))
AIC
cW
eight)
AIC
c W
eightC
EPTC
SAPL
SHR
UB
SG
FHT
FEMR
OC
KS
E73
67.3–
–0
0.3873*20.271
––
–61.295
–E6
468.2
––
1.300.2021*
!5.460
––
–61.288
–E4
467.5
––
2.130.1333*
28.118–
––
63.051!
0.329E5
467.4
––
2.340.1204*
22.717–
––
61.606–
E35
68.2–
–3.81
0.0576*!
3.080–
––
61.540–
E15
67.8–
– 4.34
0.0442*36.734
––
–64.508
!0.486
E26
68.5–
– 6.18
0.0176*10.510
––
–63.850
!0.386
C5
668.1
00.3597*
6.710.0135
34.203–
!0.0018
–64.324
!0.238
C10
667.8
0.320.3065*
7.030.0115
34.315–
––
64.430!
0.448C
17
68.22.76
0.0905*9.47
0.003414.800
–!
0.00180.737
62.669!
0.009C
47
68.12.89
0.0846*9.61
0.003234.379
–!
0.0018–
64.317!
0.240C
37
68.03.07
0.0776*9.78
0.002922.858
0.006!
0.0020–
63.092!
0.125C
27
67.93.09
0.0766*9.81
0.002911.642
––
0.93262.617
!0.144
Global
968.4
8.810.0044
15.530.0002
18.862 0.016
!0.0016
0.79962.621
!0.337
C7
618.3
39.49<0.0001
46.20<0.0001
!38.784
0.092–
––
0.519C
66
18.239.52
<0.000146.24
<0.0001 !
31.2910.096
!0.0007
––
0.456C
117
20.241.38
<0.0001 48.10
<0.0001!
56.400 0.094
!0.0014
3.463–
1.480C
87
17.742.68
<0.0001 49.40
<0.0001!
49.598 0.113
!0.0007
2.731–
0.721C
98
23.342.80
<0.0001 49.51
<0.0001226.828
0.084!
0.0003–
–!
0.267
Table 2.21. (C
ontinued).
MW
aR
2 O
riginal Models
c A
ll Models
d D
EN-
ALT-
MTM
-FW
-R
SP-M
odelK
b(%
))
AIC
cW
eight)
AIC
c W
eightLG
LOG
DEP
PREY
GO
KTEM
PR
AIN
E73
67.3–
–0
0.3873*–
––
––
–E6
468.2
––
1.300.2021*
––
––
–2.301
E44
67.5–
–2.13
0.1333*–
––
––
–E5
467.4
––
2.340.1204*
!0.121
––
––
–E3
568.2
––
3.81 0.0576*
!0.098
––
––
2.265E1
567.8
––
4.340.0442*
!0.241
––
––
–E2
668.5
––
6.180.0176*
!0.195
––
––
2.049C
56
68.10
0.3597*6.71
0.0135 !
0.189–
––
––
C10
667.8
0.320.3065*
7.030.0115
!0.252
––
0.0037–
–C
17
68.22.76
0.0905*9.47
0.0034 –
3.854–
––
–C
47
68.12.89
0.0846*9.61
0.0032 !
0.190–
!0.0001
––
–C
37
68.03.07
0.0776*9.78
0.0029–
–0.0016
––
–C
27
67.93.09
0.0766*9.81
0.0029–
3.129–
0.0017–
–G
lobal9
68.4 8.81
0.004415.53
0.0002!
0.1723.621
––
––
C7
618.3
39.49<0.0001
46.20<0.0001
––
–0.0072
–4.812
C6
618.2
39.52<0.0001
46.24<0.0001
––
––
–4.603
C11
720.2
41.38<0.0001
48.10<0.0001
0.796–
––
––
C8
717.7
42.68<0.0001
49.40<0.0001
––
––
!4.377
–C
98
23.342.80
<0.0001 49.51
<0.00010.877
––
–48.250
9.519 a Linear regression m
odels of Mexican w
oodrat abundance described in Table 2.7.b N
umber of m
odel parameters estim
ated from data including intercept coefficient, effect coefficients, and a residual error term
.c W
eighted according to minim
um A
ICc value of the 12 a priori m
odels.d W
eighted according to minim
um A
ICc value of the 12 a priori and seven exploratory m
odels.
205
Table 2.22 Precision and relative importance of ecological factors associated withsummer abundance (g/ha) of Mexican woodrats in the Sacramento Mountains, NewMexico (1992–1994). Two sets of regression coefficients ( ) are given; those averagedover a priori (original) models and the other averaged over original and exploratory (all) models. CVs of the average coefficients include variation associated with uncertainty ofmodel structure. Relative importance (RI) of each regressor variable is also shown forboth model sets and is based on the sums of Akaike weights for each model that includedthe regressor. Asterisks indicate effects included in the 95% confidence set of models. Regressor acronyms are defined in the text. Factors are ordered from highest to lowestaccording to RI in the original model set.
Original Models All Models Factor CV(%) RI CV(%) RI
MATFEM 63.972 12.7 1.000* 61.867 11.4 1.000*ROCKS !0.264 297.0 1.000* !0.082 201.1 0.233*LGLOG !0.162 200.5 0.755* !0.040 269.9 0.268*SHRUBS !0.00110 180.8 0.617* !0.0000 4 201.9 0.023MTMGOK 0.00125 604.3 0.383* 0.00005 609.2 0.014GFHT 0.142 192.2 0.172* 0.005 199.8 0.006DENDEP 0.604 356.9 0.172* 0.023 361.1 0.006ALTPREY 0.000119 1,135.2 0.162* 0.000004 1,136.4 0.006CSAPL 0.00056 571.3 0.082* 0.00002 572.6 0.003 FWTEMP .0.a 199.7 0.000 .0.a 199.7 0.000RSPRAIN .0.a 126.6 0.000 0.632 121.5 0.277* a #0.000001
206
Table 2.23. Regressor variables ranked as most important effects on summer abundance (g/ha) of five rodent species in the Sacramento Mountains, New Mexico,during 1992–1994. Two sets of effects are shown, those determined from a priori(original) models and those from original and exploratory (all) models.
Original Models All Models Species Effecta RIb CV%c Signd Effecta RIb CV%c Signd
Deer mouse MCONIF 1.000 164.1 !, ? MATFEM 0.997 23.7 +, +*PREVRAIN 1.000 329.7 !, +* FWTEMP 0.954 48.7 !, !MATFEM 0.998 23.3 +, +* GFHT 0.848 113.2 +, +*
Brush mouse MATFEM 1.000 6.8 +, +* MATFEM 1.000 7.3 +, +*ROCKS 1.000 354.6 +, +* ROCKS 0.710 299.6 !, +*SHRUBS 0.868 11,095.0 +, +* OAKS 0.436 681.8 +, +*
Mexican vole GFHT 1.000 17.7 +, +* GFHT 1.000 17.8 +, +*MATFEM 1.000 18.3 +, +* MATFEM 1.000 18.5 +, +*INFCOMP 1.000 20.5 +, +* INFCOMP 1.000 20.5 +, +*
Long-tailed MATFEM 1.000 14.9 +, +* MATFEM 1.000 13.6 +, +*vole INFCOMP 1.000 17.6 +, +* INFCOMP 1.000 17.2 +, +*
GFHT 1.000 32.6 !, +* GFHT 1.000 30.8 !, +*FWTEMP 1.000 91.3 !, vSHRUBS 1.000 748.0 +, !
Mexican MATFEM 1.000 12.7 +, +* MATFEM 1.000 11.4 +, +*woodrat ROCKS 1.000 297.0 !, v
LGLOG 0.755 200.5 !, ?SHRUBS 0.617 180.8 !, ?
a MCONIF—density ( /ha) of mature conifers; PREVRAIN—total rainfall (cm) during two previous growing
seasons (March–August); MATFEM—number of reproductively active females in the summer population;FWTEMP—mean minimum temperature (°C) during late fall and winter; GFHT—mean maximum height (cm) ofgrasses or forbs; ROCKS—mean percentage of rock cover (arcsine-square root transformed); SHRUBS—density of allshrubs; OAKS—density of shrub and tree-sized oaks; INFCOMP—density of species likely to interfere with use ofspace during the same summer; LGLOG— density of logs >30 cm in diameter at mid-point.b Relative importance based on Akaike weights of models for a given species.c Coefficient of variation of model-averaged estimates of effect coefficients.d Sign of model-averaged coefficient followed by sign of relationship in bivariate plots of indicated variable withsummer abundance of listed species; v indicates a quadratic relationship and ? indicates an unclear relationship; *denotes agreement of bivariate relationships with envirogram predictions.