Parameter adjustment in NichemapR for endotherm modeling

Pu Zhen

unread,

Sep 18, 2025, 1:34:07 AM9/18/25

to NicheMapR

Hi Professor Kearney,

My name is Zhen Pu. I am writing to seek your advice regarding why the endotherm model (endoR in NicheMapR) appears to underestimate avian evaporative water loss (EWL) at high temperatures.

I used the following code to validate the endotherm model using empirical data of Zenaida asiatica (Columbiformes) from Smith et al. (2015), where body temperature (Tb), resting metabolic rate (RMR), and EWL were measured under controlled conditions (i.e. no radiation, low wind speed, and temperatures ranging from 30–66 °C).

The body mass (AMASS) was from the AVONET and we scaled plumage depths (ZFURD and ZFURV) and feather lengths (LHAIRD and LHAIRV) from Campylorhynchus brunneicapillus in proportion to AMASS^1/3. Other non-default parameters should primarily refer to the literature or best estimates. Here is the code I used:

endo.out <- lapply(1:length(TAs),

function(x){

endoR(

#Parameters controlling how the model runs

TREGMODE = 2,

#Environment

TA = TAs[x],

#Species-specific parameters

AMASS = 0.153 #AVONET

ZFURD = 0.01418796936 # Campylorhynchus brunneicapillus = 0.009

ZFURV = 0.01103508728 # Campylorhynchus brunneicapillus = 0.007

LHAIRD = 0.04098746703 # Campylorhynchus brunneicapillus = 0.026

LHAIRV = 0.02995237975 # Campylorhynchus brunneicapillus = 0.019

REFLD = 0.25 # measurement of Campylorhynchus brunneicapillus

REFLV = 0.43 # measurement of Campylorhynchus brunneicapillus

TC = 40, # Smith et al. 2015; default = 37

TC_MAX = 45, # Smith et al. 2015; default = 39

TC_INC = (TC_MAX -TC)/1600,

#Other non-default parameters

Q10 = 1.1, #best estimate; default = 2

FURTHRMK=0.03, #Dieckmann et la. 2021

SHAPE = 4, #best estimate

SHAPE_B = 1.1, #best estimate

SHAPE_B_MAX = 3, #best estimate; default=5

PZFUR=0.002, #default=1

PCTWET = 0.05, #estimated for night parrot; Kearney et al. 2016; default = 0.5

PCTWET_MAX = 25, #best estimate; default = 100

PANT_MULT = 1.05 #default = 1.05

PANT_MAX = 10 #default = 10, #Bartholomew, G. A., Lasiewski, R. C., & Crawford, E. C. (1968). Patterns of panting and gular flutter in cormorants, pelicans, owls, and doves. The Condor, 70(1), 31-34.

PANT_INC= (PANT_MAX_spe-1)/1600,

PVEN = 0.3, #best estimate; default=0.5

SAMODE = 1, #1 is for bird skin surface area allometry from Walsberg & King. 1978. JEB 76:185–189; default=0

RHOD = 1.5e+07, #best estimate; default=3e+07

RHOV = 1.5e+07, #best estimate; default=3e+07

AK1 = 0.412, #best estimate; default=0.9

DELTAR = 5, #estimated for night parrot; Kearney et al. 2016; default=0

EXTREF = 25 #for birds; Table 39.4 in Sturkie 2012; default=20

)

The comparison between empirical and modelled results is shown in the attached figure. As you can see, the model consistently underestimates evaporative water loss at high temperatures.

I have tried adjusting parameters such as TC, TC_MAX, PANT_MULT, PANT_MAX, and Q10, but these attempts did not resolve the discrepancy. Additionally, I noticed your response to a similar issue suggested changing the PCT_WET. I also tried adjusting this parameter to the default value, but the EWL at high temperature did not show significant increases.

I wonder whether I might have overlooked some important parameter(s) or physiological process that should be incorporated into the model. I would be very grateful if you could provide some guidance on possible reasons for this underestimation, or point me toward resources that could help improve the accuracy of the predictions.

Any advice would be most appreciated!

With best regards,

Zhen

Tb EWL RMR.zip

NicheMapR

unread,

Sep 21, 2025, 5:09:09 PM9/21/25

to NicheMapR

Hi Zhen,

I played around with the parameters and got something that looks closer to your data, pasted below. Main things are to not specify FURTHRMK (leave it as 0 which is default), which overrides the calculation based on the properties of the feathers you have measured, and to alter the relative step sizes for the core temperature elevation and the panting. Your simulation had the core rising quickly compared to your observations, and with a low Q10 effect this meant the simulated bird didn't need to lose so much water.

It's worth keeping in mind that this is a steady state solution whereas in metabolic chamber experiments you may have a bird that is in more of a transient state, depending on how long you hold them at each temperature. The bird may be losing more water than it needs in anticipation of the temperature continuing to climb. You should always specify the wind speed and the relative humidity as well. Note that I specified the density of the bird as 750 rather than the default 1000 kg/m3 - see

Eichenwald, A. J., & Reed, J. M. (2023). Biased assessment of thermal properties of birds from estimated body density. Journal of Thermal Biology, 112, 103472. https://doi.org/10.1016/j.jtherbio.2023.103472

All the best,
Mike

endo.out <- lapply(1:length(TAs), function(x) {
endoR(
#Parameters controlling how the model runs
TREGMODE = 2,
#Environment
TA = TAs[x],

VEL = 0.2,
RH = 5,
#Species-specific parameters
AMASS = 0.153,
ANDENS = 750,
#AVONET
ZFURD = 0.01418796936,

# Campylorhynchus brunneicapillus = 0.009

ZFURV = 0.01103508728,

# Campylorhynchus brunneicapillus = 0.007

LHAIRD = 0.04098746703,

# Campylorhynchus brunneicapillus = 0.026

LHAIRV = 0.02995237975,

# Campylorhynchus brunneicapillus = 0.019

REFLD = 0.25,

# measurement of Campylorhynchus brunneicapillus

REFLV = 0.43,

# measurement of Campylorhynchus brunneicapillus
TC = 40,
# Smith et al. 2015; default = 37
TC_MAX = 45,
# Smith et al. 2015; default = 39

TC_INC = 0.0025, #(TC_MAX - TC) / 1600,
#Other non-default parameters
Q10 =2.5,

#best estimate; default = 2

#FURTHRMK = 0.03,

#Dieckmann et la. 2021
SHAPE = 4,
#best estimate
SHAPE_B = 1.1,
#best estimate
SHAPE_B_MAX = 3,
#best estimate; default=5
PZFUR = 0.002,
#default=1
PCTWET = 0.05,
#estimated for night parrot; Kearney et al. 2016; default = 0.5

PCTWET_MAX = 2,

#best estimate; default = 100

PANT_MULT = 1.15,
#default = 1.05
PANT_MAX = 10,

#default = 10, #Bartholomew, G. A., Lasiewski, R. C., & Crawford, E. C. (1968). Patterns of panting and gular flutter in cormorants, pelicans, owls, and doves. The Condor, 70(1), 31-34.

PANT_INC = 0.01,#(PANT_MAX_spe - 1) / 1600,

PVEN = 0.3,
#best estimate; default=0.5
SAMODE = 1,
#1 is for bird skin surface area allometry from Walsberg & King. 1978. JEB 76:185–189; default=0
RHOD = 1.5e+07,
#best estimate; default=3e+07
RHOV = 1.5e+07,
#best estimate; default=3e+07
AK1 = 0.412,
#best estimate; default=0.9
DELTAR = 5,
#estimated for night parrot; Kearney et al. 2016; default=0
EXTREF = 25 #for birds; Table 39.4 in Sturkie 2012; default=20
)

})

Pu Zhen

unread,

Sep 25, 2025, 2:17:47 AM9/25/25

to NicheMapR

Hi Mike,

Thank you very much for your valuable suggestions. I have revised the model settings following your advice, and the modelled performance has improved substantially.

Overall, the updated model works well across the majority of our focal species. However, we continue to observe a notable underestimation of evaporative water loss at high temperatures in Burhinus capensis (Charadriiformes) from Czenze et al. 2021 (https://doi.org/10.3389/fevo.2021.799302). A comparison of empirical and modelled results is shown in the attached figure.

I am wondering whether I may have overlooked one or more critical parameters or physiological processes that should be incorporated into the model. Alternatively, this discrepancy may reflect interspecific differences, and I am currently seeking published data on additional Charadriiformes species for further validation.

I would be most grateful for any guidance you could provide on potential reasons underlying this underestimation, or for any recommendations on resources that might help refine the accuracy of our predictions.

With best regards,

Zhen

endo.out <- lapply(1:length(TAs),

function(x){

endoR(

#Parameters controlling how the model runs

TREGMODE = 2,

#Environment

TA = TAs[x],

#Species-specific parameters

AMASS = 0.4340 #AVONET

ANDENS = 750, #Eichenwald, A. J., & Reed, J. M. 2023

ZFURD = 0.020077151 # Campylorhynchus brunneicapillus = 0.009

ZFURV = 0.015615562 # Campylorhynchus brunneicapillus = 0.007

LHAIRD = 0.05800066 # Campylorhynchus brunneicapillus = 0.026

LHAIRV = 0.04238510 # Campylorhynchus brunneicapillus = 0.019

REFLD = 0.2429268 # measurement of Campylorhynchus brunneicapillus

REFLV = 0.4093902 # measurement of Campylorhynchus brunneicapillus

TC = 38, # Smith et al. 2015; default = 37

TC_MAX = 43, # Smith et al. 2015; default = 39

TC_INC = (TC_MAX -TC)/800,

#Other non-default parameters

Q10 = 2.5 #best estimate; default = 2

#FURTHRMK=0.03, #Dieckmann et la. 2021

SHAPE = 4, #best estimate

SHAPE_B = 1.1, #best estimate

SHAPE_B_MAX = 3, #best estimate; default=5

PZFUR=0.002, #default=1

PCTWET = 0.05, #estimated for night parrot; Kearney et al. 2016; default = 0.5

PCTWET_MAX = 2, #best estimate; default = 100

PANT_MULT = 1.15 #default = 1.05

PANT_MAX = 10 #default = 10, #Bartholomew, G. A., Lasiewski, R. C., & Crawford, E. C. (1968). Patterns of panting and gular flutter in cormorants, pelicans, owls, and doves. The Condor, 70(1), 31-34.

PANT_INC= (PANT_MAX_spe-1)/800,

PVEN = 0.3, #best estimate; default=0.5

SAMODE = 1, #1 is for bird skin surface area allometry from Walsberg & King. 1978. JEB 76:185–189; default=0

RHOD = 1.5e+07, #best estimate; default=3e+07

RHOV = 1.5e+07, #best estimate; default=3e+07

AK1 = 0.412, #best estimate; default=0.9

DELTAR = 5, #estimated for night parrot; Kearney et al. 2016; default=0

EXTREF = 25 #for birds; Table 39.4 in Sturkie 2012; default=20

)

Tb EWL RMR from Burhinus capensis.zip

NicheMapR

unread,

Sep 25, 2025, 7:24:30 AM9/25/25

to NicheMapR

Hi Zhen,

I'm glad to hear my suggestions helped. I can't really offer much more than what I previously said. You could get in touch with Zenon who did the study and see if he has any further insights.

All the best,
Mike

Pu Zhen

unread,

Sep 25, 2025, 10:16:23 PM9/25/25

to NicheMapR

Hi Mike,

Thank you so much for your valuable suggestions.

With best regards,

Zhen

Reply all

Reply to author

Forward