Taking an AI Climate package for a test ride

Taking AI for A Climate Spin         

With so much in the news about the advances in AI ( Artificial Intelligence) I was intrigued by a post of a favorite authority about a new ChatClimate version beta version – now available for testing. So with time on my hands, here’s a look into my first “discussions” with a machine-learned software device. The discussion reflects my “human” concerns and the robot responses built upon the most recent IPCC report.

My assignment: explore ChatClimate, (CC) ask it some questions, and share your newfound knowledge with those around you. Assignment came from Dr K. Hayhoe in Talking Climate enewsletter. Dr Hayhoe is the consummate climate explainer in my book and as a longtime follower I jumped at the task. So below you’ll find the exact text of my inquiries; followed by the exact CC response altered only by my BOLD highlights of the text- short answers for conversational usage or underlining for a geographical specifics.( and author comments in italics)

Question 1: what are the recent media frequencies of reporting on climate stories since 2000

CC: According to IPCC AR6 WGIII Chapter 13, a study of 59 countries shows that global media coverage of climate change has been growing, from about 47,000 articles in 2016-17 to about 87,000 in 2020-21. The number of media science stories in those sources grew steadily from 47,376 per annum to 86,587 per annum between 2017 and 2021 across print, broadcast, digital media, and entertainment. The media representation of climate science has increased and become more accurate over time. (Page 4 and 23)

Q 2: ice free artic, which year

CC: Based on the IPCC AR6 report, the Arctic is projected to become practically ice-free in September by the end of the 21st century under SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios with high confidence. However, the report does not provide a specific year for when this will happen. The AR5 report assessed that for RCP8.5, Arctic sea ice coverage in September will drop below 1 million km2 and be practically ice-free at some point between 2040 and 2060. However, this is a medium confidence assessment based on a five-member selection of CMIP5 models. The SROCC further assessed that the probability of an ice-free Arctic in September for stabilized global warming of 1.5°C and 2.0°C is approximately 1% and 10-35%, respectively. (IPCC_AR6_WGI_Chapter04, p.22-23)

Q 3: How close are we to a methane tipping point?

CC: Reducing methane emissions is important in relation to near- and medium-term temperatures, and sustained methane mitigation offers near- and long-term gains on surface temperature and leads to air pollution benefits by reducing surface ozone level globally. The most complete modeling of these processes to date suggests a release to the atmosphere at a rate of less than 2% of current human-induced methane emissions. However, there is no evidence of a methane tipping point, and abrupt changes and tipping points in the biogeochemical cycles are very likely to be small compared to the uncertainty associated with future anthropogenic emissions. (IPCC_AR6_WGI_Chapter05, p.101; IPCC_AR6_WGIII_Chapter01, p.11; IPCC_AR6_WGI_Chapter06, p.76; IPCC_AR6_WGI_Chapter05, p.69)

Q4: Methane release in artic soils, How big a ? (Human typing and key errors are clearly not a problem for CC)

CC: There is large uncertainty in the release of GHGs from permafrost in the 21st century, but the largest estimates imply tens to hundreds of gigatons of carbon released in the form of CO2 and CH4 emissions up to 100 TgCH4 yr-1. Assuming a CH4 lifetime in the atmosphere of the order of 10 years and the associated feedback parameter, this would increase the atmospheric CH4 content by about 500 ppb over the century, corresponding to a rate of ≤10 ppb yr-1. However, irrespective of its origin, additional CH4 accumulation of such a magnitude is not expected to modify the temperature response to anthropogenic emissions by more than a few tenths of a °C. Thawing terrestrial permafrost will lead to carbon release, but there is low confidence in the timing, magnitude, and the relative roles of CO2 versus CH4 as feedback processes. CO2 release from permafrost is projected to be 3-41 PgC per 1°C of global warming by 2100, based on an ensemble of models. It is very unlikely that gas clathrates in terrestrial and subsea permafrost will lead to a detectable departure from the emissions trajectory during this century. The most complete modeling of these processes to date suggests a release to the atmosphere at a rate of less than 2% of current human-induced methane emissions. (IPCC_AR6_WGI_Chapter05, Pages 5, 56, 68, 101)    My reading: Stay tuned!

Q5: Estimate of Climate migrants through 2020 ( Human error again I tried to correct it to 2030 or 2050. CC went with what it had in the report)

CC: What is the estimated number of climate migrants by 2050? The estimated number of climate migrants by 2050 varies depending on the region and the assumptions made about future greenhouse gas emissions and socioeconomic development trajectories. In sub-Saharan Africa, with 1.7°C global warming by 2050, 17-40 million people could migrate internally, increasing to 56-86 million for 2.5°C. In South Asia, internal climate migrants are projected to be 40 million by 2050 under high warming. Model projections of population displacements by 2050 in Latin America, sub-Saharan Africa, and South Asia due to climate change vary from 31 million to 143 million people, depending on assumptions about future GHG emissions and socioeconomic development trajectories. (IPCC_AR6_WGII_Chapter06, IPCC_AR6_WGII_Chapter09, IPCC_AR6_WGII_Chapter07)

Q6: Are there human evolution correlaries to the current global temperature rise? ( yes, my typing is suspect and in this case CC didn’t correct me)

CC: The provided information does not mention any human evolution correlaries to the current global temperature rise. Therefore, there is not enough information to answer the question. ( aside from my misspelling this question was outside the scope of IPCC reports but might be a nice final exam question and I wanted to see how GPT-3.5 Turbo could/would handle it).

Take Home:

My first pass was a great success- the machine could understand my convoluted questions, spelling errors and incomplete thoughts. AI certainly can be a time saver for researchers, and anyone hoping to explore the world of IPCC reports and may only have minutes to spare versus the hours it might take to read a chapter or explore a complete volume. As with all tools, some are best utilized by those with a background in the topical area. The interpretation of the responses and “knowing” what to ask are key. Perhaps more telling is the tone and angle of the questions being asked. In my case I have viewed, read or skimmed the last two reports and had at least a cursory knowledge of their content and purpose. So while my training, education and experience are in the natural resource realms, my heart (and line of questioning) is clearly in the social impact realm. I am human, or try to be, but I am bolstered and supported by machine-learning that in time may prove to be the hope to save humanity from itself.

But please don’t take my word for it take it for a spin yourself and see what the future might be with a computer- aided search, retrieval and publication guided by a human heart with its own experience, training and scent-detection system for assessing when systems run afoul.