Will Petry

Quantitative ecologist • food webs/demography • asst prof @ncstate_pmb • fieldwork @RMBLGothic + @UMBS • he/him • stoops for charismatic microflora

[Dr. Rebecca (Becca) Tarvin 🐸 ☠️ 🐍 💚🪰 @frogsicles]

New funding mechanism from NSF for institutions to support GRFP honorable mention awardees in these states/territories: AK AR DE GU HI ID IA KS KY LA ME MS MT NE NV NH NM ND OK PR RI SC SD VI VT WV WY

Thanks @hormiga for the heads up 🧪🌿🐸$$$

https://t.co/BcovjKV24a

[Will Petry @oecodynamics]

@adam_t_clark @emlid I'd go for a single RS2+ (or RS3 to cut leveling time if you need super high precision). You can put a data SIM directly in unit or feed RTK correction from the LTE-connected device you're using to record point data

[Will Petry @oecodynamics]

@adam_t_clark @emlid I'm happy to talk more about our experience or give advice on your needs here or over email. Happy mapping! /end

[Will Petry @oecodynamics]

@adam_t_clark @emlid Be sure to look into education discount if you go this route. It's BYO phone/tablet, but iOS is much more restrictive if you want to use a non Emlid app (e.g., not paying subscription for line/polygon/survey tools). You also need a range pole. /

[Will Petry @oecodynamics]

@danielhgs829 Yes, to those with existing work authorizations. The current visa processing time in the US is very slow—up to 3-4x as long as the duration of this temporary position. Feel free to contact me by email with questions. I can connect you with our HR office if I don't know the answer

[Will Petry @oecodynamics]

Summer field tech position with my group in Colorado. I've extended the deadline to Thursday & re-posting here for anyone still around on here

Full ad & application: https://t.co/hXCwI4yti2

[Will Petry @oecodynamics]

@taddallas @ArmitageDW Bingo. Dave’s (very nice!) comparison can’t distinguish between:
•MCT is a bad theory
•the researchers did a bad job modeling their organisms
•pairwise predictions can’t capture multi species dynamics
•coexistence at larger spatial/temporal scales causes local co-occurrence

[Will Petry @oecodynamics]

@ArmitageDW There’s an asymmetry here. Big areas can encompass coexistence mechanisms operating at smaller scales, but not vice versa. Apparently paradoxical co-occurrences are just the signature of the large scale process on local pattern

[Will Petry @oecodynamics]

@ArmitageDW I think you have to restrict the coexistence mechanisms at work on the real landscape to that same spatial scale. Using my neighbors lawn again, he’s going to get weeds despite there being absolutely no way the weeds can coexist with turf grass under the current environment.

[Will Petry @oecodynamics]

@ArmitageDW I think you have a really nice contribution to make with this comparison. I’m 100% in the chorus of folks cheering you on. “MCT fails” is one end point, but there’s so much more potential in unpacking why this seemingly intuitive comparison isn’t a strong test of MCT after all

[Will Petry @oecodynamics]

@ArmitageDW Put bluntly, it’s silly to make real world predictions about coex vs excl in your species if you’re ignoring some of the most potent coex mechs. But the flip side of that coin: using co-occurrence patterns from a large, variable landscape seems like an easy straw man to defeat

[Will Petry @oecodynamics]

@taddallas @ArmitageDW Ex: My neighbor has a highly manicured lawn, while mine is…let’s say…”pollinator friendly.” If I measure ND & FD in their yard, turf would exclude the weed (b/c herbicide). Weeds would win in my yard. Competitive excl, yet co-occurrence, in both. Coexistence is at block scale.

[Will Petry @oecodynamics]

@taddallas @ArmitageDW I think spatial coexistence mechs can cause co-occurrence at local patches that are too small to allow coexistence. Parameterizing an MCT model in the local patch may predict exclusion when the stabilizing mechanisms are operating at a larger scale.

[Will Petry @oecodynamics]

@rlmcelreath @nj_clark @ali__johnston @millerdl @mcmc_stan There’s a nice Stan implementation of the ‘unmarked’ 📦 that also allows random effects. ‘stan_pcpunt()’ does the Royle 2004 N-mixture model.

https://t.co/x0N3OzAHQS