Stochastic modelling in ecosystems — progress but we are not there yet

Modelling Of Ecosystems: the Cycle, Inputs and Ouputs

Modelling Of Ecosystems: the Cycle, Inputs and Ouputs

Modelling of pro­cesses lets one under­stand the func­tions of inter­act­ing com­pon­ents, helps to identify parts of pro­cesses, and can pre­dict out­comes of changes in the sys­tem. Unfortunately, what was a major area of fin­an­cial mod­el­ling is now largely dis­cred­ited, much to the cost of the rest of us; other areas such as insur­ance are becom­ing so con­strained by rules and reg­u­la­tion as to be use­less. Biological mod­el­ling, in con­trast is advan­cing rap­idly, whether with respect to sub­cel­lu­lar events, whole organ­ism devel­op­ment, or dis­ease epi­demi­ology. This week, Professor Xueron Mao has organ­ized a meet­ing (pre­vi­ous blog post) at the University of Strathclyde in Glasgow, Scotland, on “Stochastic Modelling in Ecosystems” (link to meet­ing pro­gramme).

In the week mark­ing the 20th anniversary of the 1992 United Nations Conference on Environment and Development (“Rio Summit”), I wondered about the impact of eco­sys­tem mod­el­ling on the major policies being dis­cussed at the Rio+20 sum­mit in June 2012. Had I missed a whole area of lit­er­at­ure in the last 10 years? Weekly, the news media tell us about the res­ults from the latest mod­els of cli­mate change, while I read  papers every month about crop and pho­to­syn­thesis mod­els, not to say stun­ning work on many indi­vidual plant spe­cies, includ­ing spe­cial issues and col­lec­ted papers from Annals of Botany. Like many UN-related organ­iz­a­tions and meet­ings, the Sustainable Development con­fer­ence has a large amount of under­pin­ning ‘grey lit­er­at­ure’ – com­mis­sioned reports and research with strong sci­entific con­tent (albeit, often not com­plete or defin­it­ive, and hence not suit­able for pub­lic­a­tion in ref­er­eed journ­als). However, a search of the UN web­site does not show any attempts at eco­sys­tem mod­el­ling (or, indeed, ‘mod­el­ing’): the 14 dis­cus­sions of the topic of mod­el­ling are all about eco­nom­ics and fin­ance. The Google Scholar search shows few major papers in the last dec­ade with keywords of modelling/modeling and eco­sys­tems.

Maybe the chal­lenge of mod­el­ling a whole eco­sys­tem is too dif­fi­cult: a model needs to define inputs, out­puts and flux through a sys­tem. The eco­sys­tem involves cycles and net­works involving hun­dreds of spe­cies and mil­lions of inter­ac­tions from sub-cellular level upwards. In my own talk open­ing the meet­ing, I con­cluded that the out­puts can be clas­si­fied in three areas. Firstly, chem­ical energy, largely in the form of the fixed car­bon that is used as food, feed, fibres and fuel out­side the dir­ect eco­sys­tem. Secondly, a small but import­ant frac­tion of the flux is removed from the sys­tem, par­tic­u­larly to the long-term car­bon stores in lime­stone and fossil fuels. The final group of out­puts can be con­sidered as ‘eco­sys­tem ser­vices’ includ­ing pur­i­fied (or indeed pol­luted) water that is changed from the input state in both pur­ity and flow rate, or oxy­gen reduced from car­bon diox­ide. The slides from my talk are on Slideshare​.com under pathh, and maybe I will make a shortened com­ment­ary for YouTube at some point.

In the meet­ing, we were treated to a range of talks ran­ging from mod­els of car­bon cycles, through pop­u­la­tion and veget­a­tion dynam­ics, through to dis­ease epi­demi­ology mod­els. It is always excit­ing when dif­fer­ent research com­munit­ies come together, so it was very valu­able to hear from and talk to the math­em­aticians at the meet­ing, even if there is some dif­fer­ences in our lan­guages!
It is always invi­di­ous to pick out par­tic­u­lar talks from a full pro­gramme, and the full list­ing is given here. Since this blog is plant-related, I will note the impress­ive talks from Mathew Williams (University of Edinburgh) dis­cuss­ing how gigatons of car­bon move around the ter­restrial (and indeed atmo­spheric) car­bon cycles using global meas­ure­ments in an exper­i­ment named FLUXNET, which, along with space-based meas­ure­ments could exam­ine large-scale forest bio­mass changes over times­cales of only three years. My col­lab­or­ator Jongrae Kim (University of Glasgow http://​www​.robustlab​.org/) gave the next talk, dis­cuss­ing some formal approaches to mod­u­lar­iz­a­tion of com­plex net­works in his talk on robust­ness ana­lysis of com­munity struc­tures, of great rel­ev­ance to mak­ing very large net­works amen­able to ana­lysis. Francesco Accantino presen­ted a model of abund­ance and changes of three Acacia spe­cies in humid savan­nas adding stochasti­city to a mat­rix model, which linked nicely to Pierre Couteron (IRD, Montpellier) work­ing at other sites in sub-Saharan Africa. Pierre mod­elled the dis­tri­bu­tion pat­terns of patchy veget­a­tion, show­ing effects of rain­fall and slope in both stable sys­tems and the changes in the last 50 years. Remote sens­ing is giv­ing much more data than eco­lo­gists have ever had, and inter­est­ingly Pierre is able to use freely avail­able Google Earth for many of his ana­lyses. After valu­able talks related to aspects of epi­demi­ology in sev­eral sys­tems, the clos­ing paper by Carlo de Michele (Politecnico di Milano, Italy) built on earlier talks about water as a main determ­in­ant of veget­a­tion type – the topic of eco­hydro­logy as the study of hydro­logy that under­pins eco­logy. Like sev­eral other talks, mod­el­ling of water could give a bistable sys­tem with two solu­tions of bare soil (low rain­fall) or of veget­at­ive ground cover (high rain­fall), tak­ing into account the effects of rain­fall stochasi­city on soil water linked to veget­a­tion sys­tems. The sur­prise was that not only did the res­ults describe beha­viour of desert com­pared to top­ical forest eco­sys­tems, but also annual changes in savan­nas with dry, bare peri­ods fol­lowed by vegetation-covered wet seasons.

Stochastic mod­el­ling in eco­sys­tems” has some way to go before it becomes “Stochastic mod­el­ling of eco­sys­tems”. Genetics, meas­ure­ment meth­ods and para­met­er­iz­a­tion of prop­er­ties are com­ing from the bio­lo­gists are begin­ning to meet the mod­el­ling com­munity with their increased under­stand­ing of robust­ness, oscil­la­tion and net­work reduc­tion as well as com­pu­ta­tional approaches. I am look­ing for­ward to decisions at Rio+30 being under­pinned by recom­mend­a­tions based on rig­or­ous and robust mod­els show­ing how we can exploit eco­sys­tems without des­troy­ing the earth.

Stochastic Modelling in Ecosystems - University of Strathclyde - 2012 - Group Photograph

Editor Pat Heslop-Harrison. ORCID 0000-0002-3105-2167

Pat Heslop-Harrison is Professor of Molecular Cytogenetics and Cell Biology at the University of Leicester. He is also Chief Editor of Annals of Botany.

Pin It on Pinterest

Liked this?

Be the first to share this post with your friends!