Cet article poursuit la critique des écrits sur le climat d’Allègre et Courtillot, commencée dans la Partie I.
Prélude : C’est de la physique, idiot
…c’est bien sûr une paraphrase de la célèbre citation de Bill Clinton au sujet de l’économie. Le dernier mot est en petits caractères car nous avons appris que ce n’est pas une bonne technique de débat d’insinuer (même par inadvertance) que ceux qui ont du mal à voir la force de l’argumentation pourraient être idiots. Ce que nous souhaitons souligner par cette paraphrase est le simple fait que le lien de cause à effet attendu entre l’augmentation des gaz à effet de serre à longue vie (comme le CO2) et l’augmentation de température ne repose pas sur une vague corrélation inexpliquée entre la température et la concentration des gaz à effet de serre au cours du 20e siècle.
L’augmentation présumée de la température a été prédite bien avant d’être détectable dans l’atmosphère, en fait bien avant de savoir que le CO2 était vraiment en train d’augmenter. Ceci a été prédit pour la première fois par Arrhenius en 1896 à partir d’idées extrêmement simples sur l’équilibre radiatif, puis fut reproduit en utilisant de la physique moderne des rayonnements par Manabe et collaborateurs dans les années 60. Aucune de ces prédictions ne reposait sur des modèles de circulation générale, qui sont apparus dans les décennies suivantes et ont permis des prévisions plus détaillées. Mais la prédiction de base du réchauffement est fondée sur des principes de physique vraiment fondamentaux ayant trait à l’absorption des infrarouges par les gaz à effet de serre, à la théorie radiative du corps noir, et à la thermodynamique de l’atmosphère saturée. Chacun de ces éléments a été vérifié avec une très bonne précision par des expériences de laboratoire et des observations de terrain.
Pendant un temps, une incertitude persistait sur le fait de savoir si la vapeur d’eau allait amplifier le réchauffement avec l’amplitude simulée par les premiers modèles d’équilibre radiatif, mais une ou deux décennies de travail supplémentaire à la fois d’observation et de théorie ont montré qu’il n’y a pas vraiment de raison de douter du calcul de cette rétroaction par les modèles de circulation générale. Modifiée en introduisant l’effet refroidissant des aérosols anthropiques, la théorie rend compte de façon satisfaisante de l’allure de la variation de température des 20e et 21e siècles.
Aucune autre théorie basée sur des principes physiques quantifiables n’a pu faire de même. Si quelqu’un arrivait avec la brillante idée que, disons, le réchauffement global est dû à la pluie de Phlogistique tombant de la Lune, cela n’effacerait pas d’un coup tout ce que l’on sait sur la thermodynamique, sur l’absorption des infrarouges, l’équilibre radiatif, et la température. Au contraire, c’est le boulot de l’avocat du Phlogistique de quantifier les effets du Phlogistique sur l’équilibre radiatif, et de les ajouter de manière cohérente aux forçages climatiques existants. Quasiment toutes les tentatives d’enfoncer la théorie de l’effet de serre anthropique ont perdu de vue ce principe simple mais incontournable.
Dans un article intitulé "Are there connections between the Earth’s magnetic field and climate?" publié récemment dans le journal Earth and Planetary Science Letters, Courtillot et ses co-auteurs tentent de jeter le doute sur la responsabilité principale du CO2 dans le changement climatique récent (et probablement futur); avançant au contraire que les fluctuations du champ magnétique terrestre (en partie dues à la variabilité solaire) jouent un rôle important et négligé. Comme la plupart des travaux du genre, celui-ci est construit sur un vide intellectuel –comme si tout ce que nous savions déjà en physique du climat devait être mis au rebut pour une idée nouvelle (et en fait pas si nouvelle). Mais les problèmes ne s’arrêtent pas là. Avec l’aide d’un Commentaire publié par Bard et Delaygue (disponible ici dans EPSL ou ici comme fichier PDF), nous exposerons un ensemble d’erreurs suspectes et d’omissions qui remplissent le papier de Courtillot. Je-m’en-foutisme et ignorance, est l’interprétation de loin la plus charitable que l’on puisse apporter à cet ensemble.
Commençons par noter que, comme concurrent significatif au forçage par les gaz à effet de serre anthropiques des changements climatiques récents, le forçage radiatif directement par les variations de l’irradiance solaire arrive bon dernier. Le flux d’énergie solaire a été suivi précisément par des instruments satellitaires depuis 1978. Mesurée entre les maxima et minima des cycles solaires de 11 ans, en la moyennant sur la surface de la Terre et en prenant en compte l’albédo, l’amplitude du forçage radiatif est plus petite que 0,2 W/m2. La tendance que l’on obtient en moyennant sur plusieurs cycles solaires est encore plus faible. Ceci fait pâle figure à côté des plus de 2 W/m2 de forçage radiatif dû aux gaz à effet de serre à longue durée de vie qui se sont accumulés dans l’atmosphère depuis 1750, et encore plus par rapport au forçage qui nous attend si aucune action n’est prise pour contrôler les émissions. Rien dans la physique du climat ne suggère que la sensibilité du climat aux variations d’irradiance solaire diffère de façon substantielle de la sensibilité au forçage des infrarouges dû aux variations des gaz à effet de serre. Pour ce qui concerne le climat, un watt est (pour l’essentiel) un watt, peu importe qu’il provienne du changement de l’énergie solaire incidente ou de l’émission infrarouge due aux gaz à effet de serre.
Pour faire plus de bruit avec la variabilité solaire, il faut invoquer quelque chose d’autre pour l’impact du soleil sur le climat. Quelque chose d’exotique, comme des variations du champ magnétique. Comme il n’y a pas de mécanisme physique quantifié reliant les variations du champ au climat, Courtillot doit se rattraper en nous montrant quelques supposées corrélations entre les variations de température et du champ magnétique. Pour ne rien arranger, Courtillot n’arrive pas à décider si une augmentation du champ devrait réchauffer le climat ou le refroidir, ce qui fait qu’on ne sait même pas quelle corrélation rechercher. L’absence d’un modèle physique ne permet pas de traiter les différents forçages sur un pied d’égalité ni d’attribuer de manière fiable les causes des changements climatiques. C’est particulièrement néfaste dans le cas où les différents forçages sont fortement corrélés entre eux. Par exemple à l’échelle de l’année au siècle, la variabilité du champ magnétique, des rayons cosmiques, et de l’irradiance solaire varient pratiquement en parallèle, et donc s’il y a une corrélation avec la température (ou la couverture nuageuse) on ne peut pas dire si c’est parce que le climat répond avec une grande sensibilité directement aux variations d’irradiance, ou si quelque chose de plus exotique est en cause. Sur une période pendant laquelle la température, le forçage des gaz à effet de serre, et un quelconque index du champ magnétique augmentent tous, une technique statistique d’attribution qui ignorerait les gaz à effet de serre pour ne considérer que l’index du champ magnétique trouverait bien sûr que le champ magnétique "explique" le signal. Si nous ne connaissions rien sur l’impact climatique du CO2, ceci mettrait le champ magnétique sur un même pied d’égalité que le CO2 comme explication, mais ce n’est pas le cas. Nous en savons beaucoup sur l’impact climatique du CO2, et aucune combine avec les rayons cosmiques ou le champ magnétique ne peut faire disparaître cette physique. Ca peut même devenir encore plus confus si on oublie le rôle important des aérosols anthropiques sur le dernier siècle, comme Courtillot le fait bien trop souvent.
La confiance avec laquelle Courtillot met en doute le rôle, accepté de façon générale, du forçage anthropique sur les changements climatiques du dernier siècle est surprenante, au vu des limites fondamentales de tout argument basé seulement sur une corrélation. Mais c’est encore pire que ça : comme le montre Bard et Delaygue, la plupart des corrélations sur lesquelles reposent les pauvres cas de Courtillot et al. sont en fait bidons.
Variabilité solaire et climat : du bon, de la brute, et du truand
Les travaux sur l’influence de la variabilité solaire (et de sa proche cousine, l’influence du champ magnétique terrestre) peuvent être rangés en trois catégories. Il y a du bon, dans lequel des scientifiques consciencieux essayent de débrouiller avec leur meilleure objectivité un signal complexe et probablement faible (mais cependant important). Comme exemples de travaux de cette catégorie, je citerais les efforts inlassables de Judith Lean pour relier l’irradiance au nombre de taches solaires; les travaux de Bard et ses collègues sur le développement d’indicateurs isotopiques du soleil comme le 10Be; les travaux de Shindell sur la réponse à la variabilité des ultraviolets; et les travaux de Foukal et al. sur les facteurs de variation de l’irradiance solaire. J’ajouterais aussi le travail récent de Camp et Tung de diagnostic de l’amplitude du cycle solaire dans la température dans la catégorie du "bon" – que ce soit un papier facile à détourner par les sceptiques de l’effet de serre n’enlève rien à sa qualité scientifique. En fait, je dirais que la plupart des travaux sur le sujet du climat et de la variabilité solaire relèvent de la catégorie du bon. C’est plutôt bien. En fait, les scientifiques ont depuis longtemps reconnu l’importance de la variabilité solaire comme l’un des facteurs gouvernant le climat (voir la synthèse très académique sur ce sujet de Bard et Frank, accessible ici à EPSL et ici en PDF). La connaissance de la variabilité solaire doit être (et elle l’est) prise en compte pour expliquer le changement du climat du dernier siècle, et pour tenter d’estimer la sensibilité climatique à partir des variations récentes du climat. De plus, le Petit Age Glaciaire nécessite une explication, et la variabilité solaire représente actuellement la seule possibilité fiable (il est moins clair que l’Optimum Médiéval soit un phénomène suffisamment cohérent pour nécessiter une telle explication).
Et puis, il y a de la brute, avec comme exemples typiques deux papiers de Scaffetta et West discutés sur RealClimate ici et ici. C’est juste de la science ‘normalement’ mauvaise, dans le sens que l’approche suivie par les auteurs est erronée quelque part, ce qui aboutit à des conclusions fausses. Peut-être qu’une partie de ce travail n’aurait jamais dû passer à travers la critique des pairs, mais du moment que les méthodes sont bien expliquées et honnêtement décrites, les investigateurs suivants seront capables d’identifier les erreurs et donc de confirmer ou d’infirmer les résultats.
Et puis… il y a du truand. Ces papiers franchissent la ligne séparant l’erreur simple de la tromperie active. Les papiers de cette catégorie commettent ce que Damon et Laut nomment judicieusement un "ensemble d’étranges erreurs". Les papiers de cette catégorie utilisent souvent des manipulations de données criticables (et souvent cachées et non expliquées) pour fabriquer des corrélations quand elles n’existent pas. Les travaux des fanatiques danois du soleil, largement discutés par Damon et Laut, sont des exemples typiques de cette catégorie de truand. Nous laissons au lecteur le soin de décider, après la discussion qui suit, si le papier de Courtillot est simplement du brut, ou s’il a versé dans le truand.
Manipulation contre académisme
Le style général du discours de Courtillot et al. a plus à voir avec le genre polémique et partial que l’on trouve chez Lomborg ou dans l’article simulant un PNAS de Robinson et al., distribué avec le projet original de pétition contre le réchauffement global, qu’il a à voir avec le genre académique qui essaye d’approcher la vérité. Il cite des papiers sans aucun sens critique et seulement s’ils peuvent étayer la thèse de l’auteur (comme par exemple l’utilisation non critique du papier biaisé de Scafetta et West cité précédemment pour appuyer une attribution importante du changement climatique du 20e siècle à la variabilité solaire). Il y a aussi pas mal de pagaille, par exemple les gaz à effet de serre sont listés en derniers parmi d’autres causes de changement climatique, sans indication de l’amplitude relative des différents forçages. Entr’autres problèmes on peut citer :
- Courtillot exagère le forçage radiatif des nuages d’un facteur quatre, car il attribue pratiquement tout l’albédo terrestre aux nuages et oublie de prendre en compte l’effet de serre des nuages.
- Il dit que "le refroidissement entre 1940 et 1970 est souvent négligé comme faisant partie du bruit" alors qu’en fait c’est l’étude approfondie de cette période qui amena les scientifiques à reconnaître l’importance de l’effet des aérosols anthropiques, à l’époque du second rapport d’évaluation du GIEC. En ignorant encore une fois l’importance bien documentée des aérosols anthropiques, il dit plus loin : "Notons que la stabilisation ou diminution de la température entre 1940 et 1970 correspond bien aux séries solaire et magnétiques, au contraire de l’accélération monotone de la hausse du CO2". Non seulement, c’est une dissimulation à la Crichton d’un phénomène bien connu, mais comme nous le verrons plus loin la supposée "correspondance" est le résultat d’une manipulation discutable des données.
- Courtillot indique une étude de Crowley utilisant un modèle d’équilibre énergétique pour étayer sa thèse qu’il manquerait de la physique dans ces modèles, ce qui affecterait la réponse au forçage solaire. Précisément, Courtillot pointe un désaccord entre modèles et données au début du 20e siècle. Toutefois, Crowley n’a pas inclus l’effet indirect des aérosols, et un modèle d’équilibre énergétique n’a pas de géographie et on ne peut donc s’attendre à ce qu’il simule fidèlement des caractéristiques comme la différence du cycle saisonnier entre continent et océan ou la couverture de glace et de neige. Les modèles de circulation générale forcés avec une combinaison de forçages naturels (y compris le solaire) et anthropiques (gaz à effet de serre et aérosols) n’ont aucun problème pour reproduire le climat du début du 20e siècle. De plus, le modèle utilisé par Crowley simule précisément la réponse climatique au forçage solaire plus tôt dans le millénaire, il est donc difficile de voir pourquoi la "physique manquante" devrait seulement se faire sentir en 1850. Il est toujours suspect lorsque des citations isolées sont utilisées pour tirer une conclusion exactement opposée à celle de l’auteur même du papier. Pour rappel , voici ce que Crowley lui-même dit dans ce papier de ses propres résultats :
- Il y a ainsi deux arguments indépendants de la nature anormale des températures de la fin du 20e siècle. Premièrement, le réchauffement du dernier siècle est sans précédent sur le dernier millénaire. Deuxièmement, ce modèle capable de simuler une bonne partie de la variabilité de la température de l’hémisphère nord sur la période 1000-1850 indique que seulement 25% environ du réchauffement du 20e siècle peuvent être attribués à la variabilité naturelle. L’essentiel du réchauffement du 20e siècle est cohérent avec celui prédit par l’augmentation des gaz à effet de serre. Ces deux arguments supportent de plus l’idée que l’effet de serre anthropique est déjà là.
…et maintenant pour la partie de vrai truand
Bard et Delaygue ont mis à jour un certain nombre d’erreurs d’une nature plus troublante. Courtillot et al. commettent l’erreur de la "Terre plate" d’où provient le titre de notre article : ils donnent une impression trompeuse de la comparaison entre les forçages dû à la variabilité solaire et dû aux gaz à effet de serre, en oubliant de prendre en compte la géométrie sphérique de la Terre et son albédo. Après l’humiliation publique endurée par Le Mouël à ce sujet pendant les débats de l’Académie (voir la partie 1), dans son article paru dans La Lettre Courtillot se donna beaucoup de mal pour montrer qu’il avait bien compris les conséquences de la rotondité de la Terre. Pourtant, cette compréhension nouvelle n’a donné lieu à aucun signe de corrigé envoyé à EPSL, ce qui nous oblige à conclure que la tromperie est délibérée. De plus, dans leur Figure 1, Courtillot et al. montrent des données géochimiques provenant d’une stalagmite des Alpes Centrales, données qui prétendent établir une corrélation très forte entre des variations climatiques et un indicateur de l’activité solaire. Comme le notent Bard et Delaygue, Courtillot et collaborateurs ont caché le fait que la corrélation est aussi bonne précisément parce que la chronologie des séries comparées a été finement ajustée afin de maximiser expressément la corrélation. Les données originales non ajustées ne montrent pas une si bonne corrélation.
La pièce de résistance de Courtillot et al. est le graphe suivant, qui prétend montrer que durant presque tout le siècle dernier la température a été fortement corrélée avec l’activité solaire et la variabilité du champ magnétique. Les trois courbes du graphe sont, d’après le papier, l’enregistrement de la température moyenne globale de Phil Jones (Tglobe, ronds rouges), une reconstitution de l’irradiance solaire totale (S(t), triangles roses; Courtillot cite la reconstitution de Solanki dans le texte), et un index de la variabilité du champ magnétique pour un site en Ecosse (ESK, en bleu) et le site de Sitka en Alaska (SIT, en vert). Les courbes ont été centrées pour avoir les mêmes moyenne et déviation standard sur leur période d’enregistrement afin qu’elles soient plus proches. Notons que la courbe S(t) s’étend sur une période plus courte que les autres; ceci aura son importance.
Impressionant, hein? Et bien cela le serait aux petits détails près que "S(t)" n’est en fait pas l’irradiance solaire totale, "Tglobe" n’est en fait pas la température moyenne globale de Phil Jones comme annoncé, et ni "ESK" ni "SIT" ne ressemblent vraiment aux indices plus larges géographiquement de variabilité magnétique qui sont des indicateurs plus fiables de l’activité solaire. Bard et Delaygue ont trouvé curieux que Courtillot n’utilise que la toute fin de la reconstitution de Solanki alors que le siècle entier est disponible. Ils ont vérifié ce que la courbe donnerait si elle était normalisée sur toute sa longueur. C’est la courbe grise en gras dans leur version corrigée de la figure ci-dessous. Pour comparaison, la courbe violette avec des triangles correspond à la reconstitution de Solanki tronquée sur la même période que celle choisie par Courtillot.
Vous saisissez ? En découpant le dernier bout de la courbe et en le normalisant à une déviation standard, Courtillot gonfle la variabilité et rend la corrélation meilleure qu’elle ne le serait en utilisant la totalité de la courbe. Comme morceau de manipulation trompeuse de données, celui-ci doit rejoindre dans les tréfonds de l’Histoire le lissage sélectif utilisé pour certains enregistrements solaires, comme discuté par Damon et Laut dans leur critique des travaux danois sur le soleil. Maintenant, dans sa réponse à Bard et Delaygue (il y a toujours une Réponse à un Commentaire), Courtillot s’enfonce encore un peu plus. Il déclare que la raison pour laquelle il a utilisé une série solaire tronquée est que les données ne proviennent pas de Solanki (comme suggéré dans le papier) mais en fait de sorties du modèle SOLAR2000 de Tobiska. Le papier de Tobiska n’est même pas cité par Courtillot et al. (2007), alors que Solanki (2002) y est cité comme dans les papiers précédents de ces auteurs sur des sujets similaires. Il n’y a aucune raison valable pour utiliser SOLAR2000 dans une étude comme celle tentée par Courtillot et al. car, comme le notent Bard et Delaygue, le modèle SOLAR2000 est limité à la petite fraction ultraviolette du spectre solaire, ce qui n’en fait pas le bon choix, sauf si on s’intéresse explicitement à des phénomènes liés au forçage des ultraviolets (voir Lean, 2002). On peut penser que Courtillot et al. sont allés pêcher ces données dans la mare la plus proche, parce que c’était la première courbe trouvée qui leur donnait quelque excuse pour tronquer l’enregistrement de façon à obtenir le résultat voulu.
Bard et Delaygue ont remarqué autre chose d’étrange. La courbe "Tglobe" de Courtillot ne ressemble pas vraiment à celle publiée par Phil Jones. La courbe de Phil Jones, provenant de ses vrais fichiers de données, est représentée dans la version corrigée par Bard et Delaygue de la figure. Ils ont également ajouté la reconstitution de la NASA pour comparer. Ces deux courbes sont en accord, mais aucune ne montre la succession rapide de réchauffement/refroidissement entre 1940 et 1970 visible sur la figure de Courtillot. Alors si les données de Courtillot ne sont pas les températures moyennes globales de Phil Jones, qu’est-ce-que Courtillot a représenté ? Nous ne le saurons peut-être jamais. Dans sa réponse à Bard et Delaygue, Courtillot assure que ses données proviennent d’un fichier nommé : monthly_land_and_ocean_90S_90N_df_1901-2001mean_dat.txt. Bard et Delaygue soulignent cependant que Phil Jones n’a connaissance d’aucun fichier de la sorte dans sa base de données, et ne reconnait dans la supposée courbe "Tglobe" aucune version de la température moyenne globale que son groupe ait pu produire.
Quelle que soit l’origine des supposées données de "Tglobe" utilisées dans le papier de Courtillot et al., il n’y a aucune raison valable -dans un papier publié en 2007- pour tronquer la série de températures en 1992 comme ils l’ont fait. Il y a, cependant, une très bonne raison illégitime, dans le fait que tronquer cette courbe de la sorte aide à cacher l’amplitude de la tendance au lecteur, ainsi qu’à raccourcir la période pendant laquelle la discordance entre l’activité solaire et la température est la plus flagrante.
Dans le graphe corrigé, Bard et Delaygue représentent également l’index géomagnétique "aa". C’est un index basé sur deux stations situées aux antipodes, index qui se trouve bien corrélé avec la variabilité géomagnétique globale estimée à partir d’un réseau plus large de stations. On peut montrer que si on s’intéresse aux températures moyennes globales, l’index aa représente une meilleure base de comparaison que les enregistrements de station unique située à haute latitude nord utilisés par Courtillot. Notez que l’index aa suit bien l’irradiance solaire de Solanki, à la différence des mesures de station unique.
Dans le graphe corrigé, Tglobe, aa et S(t) se suivent en augmentant de 1900 à 1940, mais notez que les gaz à effet de serre augmentent aussi de façon monotone sur cette période, et après bien sûr. Une attribution purement statistique pourrait imputer presque tous les changements entre 1900 et 1940 à la variabilité solaire ou magnétique, mais la même technique pourrait les imputer aux gaz à effet de serre. Seule la physique permet de trancher sur cette responsabilité. Cependant, depuis 1940, il n’y a pas la moindre apparence de corrélation entre Tglobe et S(t) ni avec aucun des indices magnétiques. L’index solaire et l’index aa présentent tous deux une bosse vers 1950, alors que la température est constante ou en diminution. L’analyse erronée de Courtillot repousse, elle, l’absence de corrélation jusqu’en 1985.
Entre le ‘show’ embarrassant lors des débats à l’Académie et le pastiche de science exposé par Bard et Delaygue avec le papier d’EPSL, vous penseriez que Courtillot serait à la recherche du plus proche trou de souris pour s’y cacher. Bien loin de cela, il a été aperçu récemment donnant une présentation au titre de "Quel réchauffement climatique ?" à cet événement prestigieux regroupant de nombreux physiciens et chimistes de renom. Certaines personnes ne connaissent pas la honte.
Postlude : de l’or et du plomb
Bard et Delaygue concluent avec une figure, reproduite ci-dessous, qui illustre bien ce que nous disons depuis des années à RealClimate. Sur cette figure ils ont représenté la température moyenne globale de Phil Jones avec un index global magnétique (l’index aa), un index du flux de rayons cosmiques (Climax), et la série composite PMOD des mesures satellitaires de l’irradiance solaire. Ces courbes sont moins lissées que celles du graphe précédent. La variabilité inter-annuelle de la température est liée à des causes naturelles comme les éruptions volcaniques majeures, les événements ENSO, et la variabilité solaire. Cependant, seule la courbe Tglobe est caractérisée par une tendance à l’augmentation très significative -une tendance qui ne peut être expliquée par ces causes naturelles.
Répétez-le trois fois chaque soir avant d’aller vous coucher : la température augmente. L’influence solaire monte et descend, monte et descend, monte et descend. Vous ne pouvez pas plus transformer ça en une tendance que faire de l’or à partir du plomb.
Une note ajoutée aux épreuves, concernant la "Note ajoutée aux épreuves"
La discussion ci-dessus était basée sur les versions du Commentaire de Bard et Delaygue et de la Réponse de Courtillot telles qu’elles étaient disponibles sur le site Internet de Elsevier jusqu’au 15 décembre. Depuis le moment de son écriture, des changements étranges ont eu lieu sous la direction de l’Editeur responsable, Robert van der Hilst du MIT. Il a effacé la "Note ajoutée aux épreuves" de la version finale du Commentaire de Bard et Delaygue. Bard et Delaygue ne l’ont découvert qu’en recevant les épreuves de leur Commentaire. Ce qui est encore plus dérangeant est que van der Hilst a autorisé Courtillot à modifier le texte de sa Réponse en se basant sur ce que Bard et Delaygue ont écrit dans la "Note ajoutée aux épreuves" maintenant effacée. Bard et Delaygue n’ont eu aucune possibilité de voir ou de commenter ces changements. J’ai laissé la discussion ci-dessus comme elle l’était, afin que le lecteur puisse se faire une meilleure opinion de cette étrange histoire de succession commentaire/réponse.
Dans la version révisée de sa "Réponse", Courtillot admet maintenant que la série de température appelée "Tglobe" ne provient d’aucune base de données de Phil Jones. Courtillot déclare maintenant que ses données proviennent d’une étude de Briffa et al. (2001), en indiquant l’adresse sur Internet du fichier entreposé au NCDC. Comme spécifié dans cette étude ainsi qu’en entête du fichier, les données ont été “recalibrées pour estimer les températures moyennes d’avril à septembre pour les régions continentales situées au nord de 20°N”. Les données de température utilisées par Courtillot ne correspondent donc absolument pas à Tglobe, ne représentent pas l’hémisphère entière, et de plus ce ne sont même pas des moyennes annuelles.
Rob van der Hilst — ancien visiteur de l’IPGP (l’institution de Courtillot)– assure que ces modifications ont été faites dans l’intérêt de la communication scientifique. Je laisse au lecteur juger si ces actions ont été appropriées, ou si elles n’ont été qu’une tentative pour tirer Courtillot de l’embarras. Dans l’intérêt de la communication scientifique, j’annexe ci-dessous le texte entier de la "Note ajoutée aux épreuves" qui a été enlevée du Commentaire de Bard et Delaygue :
Note ajoutée aux épreuves :
Dans leur Réponse à notre Commentaire, Courtillot et al. déclarent que pour la courbe d’irradiance totale S(t) ils ont utilisé le produit du modèle SOLAR2000 de Tobiska (2001) au lieu de la série de plus d’un siècle de Solanki (2002) citée dans leur papier original (Courtillot et al. 2007). Cependant, le modèle SOLAR2000 est restreint à la composante UV et leur irradiance solaire totale est fortement biaisée, comme souligné par Lean (2002).
Pour la courbe de température globale Tglobe, citée comme étant de Jones et al. (1999) dans Courtillot et al. (2007), ces derniers déclarent maintenant dans leur réponse qu’ils ont utilisé le fichier de données suivant : monthly_land_and_ocean_90S_90N_df_19012001mean_dat.txt. Nous n’avons pu trouver ce fichier, même en contactant son auteur présumé qui nous a spécifiquement déclaré qu’il ne s’agissait pas de l’un de ses fichiers (Dr. Philip D. Jones, communication écrite datée du 23 oct. 2007).
Tobiska, W. K. 2001, Validating the solar EUV proxy, E10.7, J. Geophys. Res. 106, 29,969-29,978.
Lean, J.L., 2002. Comment on ”Validating the solar EUV proxy, E10.7” by W. K. Tobiska. J. Geophys. Res. 107, (A2), 1027, DOI: 10.1029/2001JA000137.
The newscientist had an article that seemed to blame the fall in global temperatures in the 50s to sulphate aerosols.
http://environment.newscientist.com/channel/earth/mg19225822.300-histories-the-ice-age-that-never-was.html
[Response: Yes, indeed. That is the best-supported explanation for the interruption in global warming in the 50’s, and indeed in some regions this even shows up as an actual decline in temperatures. There is more uncertainty in aerosol forcing than in GHG radiative forcing, but what I said about “it’s the physics” applies to cooling influences like aerosols as well as warming influences like GHG — if somebody wants to say phlogiston (or magnetic fields) explain mid-century cooling, they can’t just ignore the known influences of aerosols when trying to do an attribution. –raypierre]
raypierre
for the Camp and Tung paper, I don’t understand, as I said in a precedent post, how it is possible that a 0.17W/m2 radiative forcing leads to 0.18°K temperature increasing.
The 0.17W/m2 solar forcing in 5.5 years (between a minima and a maxima) corresponds to 0.031W/m2.y .
You said that oceanic thermal inertia doesn’t react in the same manner because it is a rapid variation compare to GH effect increase.
For this last effect,if I take +2ppmCO2/y I find 0.028W/m2.y.(for CO2 only and without other GHG, aerosols, ..).
So how do you look at this (apparent for me) contradiction?
[Response: You’ve put your finger on precisely the puzzling thing about the C&T results, which will need to be sorted out by a more detailed analysis of the oceanic response in the IPCC AR4 models. I can only explain their result by a combination of high sensitivity and low thermal inertia. But why should the thermal inertia that explains the solar cycle response be so much lower than the thermal inertia needed to explain the seasonal cycle over oceans? Camp and Tung are on to something interesting, but I don’t think it can be said that we understand what is going on yet. ]
But I think that also part of your problem is that you double-counted the geometric factor. That .17 W/m**2 I quoted already has the albedo and sphericity terms taken into account. As for your comparison to CO2, it is indeed true that taken over the time between solar min and solar max, the solar radiative forcing fluctuation is larger than that due to CO2 increase over the same period. However, when you compare the min to max solar forcing to the net GHG radiative forcing over the past century, solar looks like a minor player — especially in view of the fact that the min to max solar cycle amplitude is not the relevant number, For that , one needs to look to the long term trends in solar irradiance. –raypierre]
Raypierre mentioned phlogiston, clearing scoffing at the likelihood that it has anything to do with climate. However, last summer, in a journal RC readers might recognize, I ran across:
“Climate science and the phlogiston theory: weighing the evidence”
Arthur Rorsch, Energy & Environment 18, number 3-4, 2007. p441-447(7).
http://www.ingentaconnect.com/content/mscp/ene/2007/00000018/f0020003;jsessionid=7abb5un7rjvfi.henrietta
But, seriously, thanks for the nice analysis.
Great deconstruction job, Ray.
C’est accablant , but alas, totally unsurprising. Here is a little tale to give it context.
I happened to be passing through the IPGP this past January, when Courtillot and Allègre launched “allègrement” their Spring 2007 “climate skepticism seminar series”. It started with the visit of well-known climate skeptic Richard Lindzen (MIT) – one of the few skeptics to actually know something about climate .
What non-French readers may not be aware of, is how abrasively obnoxious Allegre and his clique have been to the country and the entire French research community – some will say since 1968, when Allegre apparently showed himself so unbearable that he got literaly thrown out of the window at a student meeting in La Sorbonne (first floor …). There are countless such stories which need not be repeated here.
As a young and innocent student who got into Earth Sciences in 1998 (while Allègre was still ministre de la Recherche et de l’Education ), all i could hear about, anywhere i went, was resentment towards his bullish and pretentious practices. I originally thought it was born out of jealousy for his political success, but had to surrender to the evidence : obnoxious is his style and did not improve after his tremendous unpopularity forced him to resign from the ministry.
On the contrary, it seems that he’s been missing media attention , and that gratuitous climate skepticism has given him an ideal platform to get closer to a mic.
The Lindzen seminar started by a pompous introduction by his first lieutenant (squire ?) Courtillot about how misguided climate science currently is, and the need to do this work “properly”. His tone was the usual heatedness and anger that i have seen stain all his recent public addresses. Needless to say, there was considerable uproar in the room, with much of the Parisian climate community come to hear the fireworks. Allègre’s clique on the front row vs climate scientists filling much of the room behind them. Lindzen bemusedly watched the verbal joust for a good 10 minutes, fully aware that he the fight was not about him.
Clearly, it was not about climate science : it was about Courtillot and Allègre throwing “a stone in the pond” of climate scientists, desperately trying to come out of political oblivion.
Add to that Courtillot’s notorious tendency of defending ideas in the face of a towering pile of evidence (like that of all major mass extinctions, in particular the Cretaceous/Tertiary one, having been caused SOLELY by volcanic eruptions), and there is little wonder what this is about : Devil’s advocacy as an attention-seeking practice.
It is very sad, because their track record does prove that they can do good science when they want to. Like Lindzen…
I admit I have my doubts about the Camp & Tung result. I’m not saying it’s wrong, just that I have my doubts.
For one thing, they detect solar-cycle response of 0.18 K, but in an earlier version of their paper the same graph which has this number has a different number quoted, namely 0.10 K — I’d need more details about their procedures to understand why. For another thing, I’m doubtful of the validity of their projection of the global temperature spatial pattern onto the “composite mean difference,” by which they transform global average temperature to a newer, modified temperature time series — I suspect this step may artificially inflate the solar-cycle response. Again, I’d need more details to say with confidence. And then there’s the fact that none of the temperature time series I’ve studied (HadCRU-T3 and GISS GLB_Ts+SST) show statistically significant response to the solar cycle, although they do not constrain the response to be less than the Camp & Tung value (at the upper limit of the error range).
If Camp & Tung turn out to be right, it wouldn’t disturb me too much in terms of the rapid response. After all, the climate system has more than one active time scale; while the oceans drag their feet, the land-atmosphere system responds quickly (as volcanic eruptions confirm).
In any case, I repeat that I’m not saying they’re wrong, just that I have my doubts.
As for the paper by Courtillot which is the subject of this post, it looks to me like it has gone way past “bad,” deep into the bowels of “ugly.”
Raypierre,
What an interesting post. It reads something like a spy novel.
The ‘Note added in proof’ appears in the PDF version at the link you gave: https://www.realclimate.org/images/BardDelaygue.pdf. That version is the ‘Article in Press’ version, not the published version. You may want to clarify this to avoid confusion.
As you say, the ‘Note added in Proof’ is gone from the online version currently available at on the website at ScienceDirect. It is intriguing to compare the ‘Article in Press’ version with the published version. Very strange editorial practice.
[Response: Yes, I ought to clarify that. The pdf linked is the version I downloaded from the EPSL site before van der Hilst deleted the Note. I made a separate copy mainly so readers without access to the journal could still read the comment, but in the end it served a separate purpose in providing a record of the original form. –raypierre]
Thanks for mentioning phlogiston, one of my favourite examples of how scientific theories develop.
This whole episode reminds me of one of my observations on the overall debate. It is a sophomoric debating trick to bolster a weak position by accusing the other side of all the flaws in your own argument — very effective when being judged by an audience unable to evaluate the argument on its merits.
John Mashey’s reference (#3) to a phlogiston paper provides a classic example. Almost everything they say to establish conventional climate science as the new phlogiston applies much more aptly to their side of the debate.
Let’s try a few arguments from the contrarian side:
* the “mainstream” is about correlations with no underlying physical model
— nope, there are numerous examples of this kind of bogus study in the contrarian camp, especially the various “blame the sun” positions which lack elementary tests of physical validity, like does the solar variance measured account for the change in observed temperature?
* the fits look good because the data has been massaged
— again, the solar crew is widely guilty of this: see not only this paper but also Soon and Baliunas’s “it’s all the sun” papers
* the mainstream is ignoring significant factors, only looking at CO_2
— not guilty: many contrarians on the other hand are happy to model the entire climate as controlled by a single variable; IPCC et al. generally try to include all quantifiable influences
I could go on but you get the picture.
The point is to recognize this stuff for what it is, and expose this debating tactic.
Great post, Raypierre! My you’ve been busy the past week or two!
[Response: Yah, time to get back to writing the book, I guess. At least I still managed to finish grading my final exams only a week late. I very much appreciate all the words of encouragment from the readers. –raypierre]
Minor quibble: Arrhenius 1896 provided the first measurement of the radiative forcing due to CO2 and estimated the potential impact of variations in its concentration. However, it wasn’t until ~1908 that Arrhenius connected this to fossil fuel consumption to predict that temperatures would increase. Before that realization, speculations about CO2 were purely of academic/historical interest.
Incidentally, he estimated that it would take 3000 years for fossil fuel consumption to have a meaningful impact on climate. Sadly, he underestimated man’s enthusiasm for burning fossil fuels.
You forget two things involving Allegre:
1. On the timescale of the Archean, a century and 10,000 years are both negligible.
2. A change in mean surface temperature of 5 degrees make no substantial difference to the cooling curve of a 1750 degree komatiite.
Physicists have put so much effort into measuring absorbtion and reflection of every wavelength by every concievable form of matter. There are also innumerable spectrometers in chemistry departments. I haven’t checked, but I would guess that other departments would also have spectrometers. And you would think that Courtillot did not have access to the MIT Wavelength Tables.
It is really hard to understand how anybody could avoid the idea that the absorbtion bands of every gas have been cataloged and re-measured about a jillion times. Perhaps all high school students should be required to take 4 years of physics in which they spend about a semester measuring the spectra of CO2 and other things.
Another subject I would like to read RealClimate’s comments on: See the chart on page 274 of “Six Degrees” by Mark Lynas. Mark Lynas says we have until 2015 to BEGIN REDUCING our total CO2 output and we have until 2050 to actually reduce our CO2 output by 90%. Mark Lynas says if we don’t follow the schedule in Six Degrees, we will encounter positive feedbacks which will take the control of the climate out of our hands. Civilization may fall anyway well before 2050, but we can avoid going extinct by 2100. Mark Lynas says we have to hold the CO2 level to 400 parts per million to have a 75% chance of avoiding the positive feedbacks. Is Mark Lynas correct? 8 years is a very tough timetable to stop the building of coal fired power plants and replace some coal fired power plants with nuclear. I doubt that anything else other than a plague that kills a few billion people could make a dent within 8 years.
[Response: From other estimates I’ve seen, Lynas’ timetable seems about right if the goal is to avoid 450 ppm. To avoid 400ppm, even his timetable is a bit of a stretch. However, with regard to the impacts of exceeding 400ppm (or even 450ppm), if you are quoting Lynas correctly I would differ with his assessment. There is no magic threshold crossed at 450 that commits us suddenly to the kind of catastrophic changes you mention, and certainly not to extinction of humanity by 2100. If we can’t hold the line at 450, there are still harms to be avoided by stopping short of doubling. If we can’t stop doubling, there are still harms to be avoided by preventing tripling, and so forth. But his general sentiment that we can’t drag our feet on this is correct. –raypierre]
re4
Thanks for reminding us of the deposit of academic skeletons at the K-T boundary. America may require automation to catch up to the reflexive polemics of the IPGP:
http://adamant.typepad.com/seitz/2007/02/climate_of_list.html
Interesting.
One thing I’m very confused about. I assume what we are talking about here is not inherent variations in the Earth’s magnetic field (e.g. its dipole moment) due to the Earth’s internal dynamo, but rather variations in the external field due to external forcings (e.g. the solar wind)? What is the basic idea? Is that energy is pumped into the atmosphere, say, by Alfven waves coming down through the ionosphere?
It would be nice to have an explanation of the physical mechanism that Courtillot supposes, and a back-of-the-envelope calculation giving the order of magnitude of the effect he would expect based on theory, independently of the empirical results. That’s the sort of thing I’d expect as a referee, anyway.
Thanks,
Peter
> It is really hard to understand how anybody could avoid the
> idea that the absorbtion bands of every gas have been cataloged
> and re-measured about a jillion times.
Those avoiding this utter certainty that the science knows everything precisely can mostly be found publishing papers in the field.
I won’t pretend I know what this means, but I take it they would disagree with your claim that everything is known already:
“The following improvements to Heller’s theory have been made: (a) derivation of new recurrence relations for the time-dependent wave packet overlap in the case of frequency changes between the ground and electronically excited states, (b) a new series expansion that gives insight into the nature of Savin’s preresonance approximation, (c) incorporation of inhomogeneous broadening effects into the formalism at no additional computational cost, and (d) derivation of a new and simple short-time dynamics based equation for the Stokes shift that remains valid in the case of partially resolved vibrational structure. …”
http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=JCPSA6000127000016164319000001&idtype=cvips&gifs=yes
J. Chem. Phys. 127, 164319 (2007) (15 pages)
Published 30 October 2007
[Response: There’s plenty in spectroscopy remaining to be understood, particularly with regard to the CO2 continuum as it applies to Venus and Early Mars. The theory behind collision-induced absorption is still pretty flaky. None of that compromises the kind of stuff that’s needed to accurately calculate the GHG radiative effect in Earthlike conditions. ]
Minor point but your lead quote is from James Carville and not Bill Clinton.
[Response: Thanks for that. Though Carville coined the phrase, it was widely used in the 1992 Clinton campaign; I can’t say that I’ve found any evidence that Clinton himself ever uttered it, though. Digging around, I also found out that the sign Carville posted in headquarter’s didn’t have the “it’s” which somehow got added in later. –raypierre]
Raypierre : the SOLAR2000 model is restricted to the ultraviolet portion of the solar spectrum
I used to think SOLAR2000 model (now Solar Irradiance Platform – SIP) is an irradiance specification tool which produces the full solar spectrum of historical Sun variations (or spectral formats, like UV, if needed for research). Do you have more precise information for SOLAR2000 limited to UV spectrum?
[Response: What I said applies more specifically to the Tobiska et al. application. Take a look at the paper by Lean cited in the “Note Added in Proof” and see if that helps. –raypierre]
Interesting case, Ray. Media and the public are constantly remarked to refer to peer reviewing as the only credible source of information. The EPSL is not the last scientific publication, it has a reputation. If this case has been unveiled, how many other cases of “strange” reviewing may have been passed without notice?
And should the public be informed about such cases, or may this harm science at all?
[Response: Certainly the public should be informed. Cover-ups are never good. When faults in the editorial process are exposed, there’s a better chance that they will be fixed. It would be easy to misread an incident like this as devaluing everything in the peer-reviewed literature, but that would be unfair. Cases like this are pretty rare. Usually, reviewing works pretty well, and even when something bad (or ugly) slips through, it is caught by the scientific community pretty quickly, if the result is important enough to check. That’s why it’s important not to make judgements based on an individual paper, but instead to watch for the follow-up. –raypierre]
Re #11, yes he is essentially correct I believe. The subject has been discussed here under the term “alarmist and alarmism”. Due to uncertainties and non linear nature of the earth system of which climate is a part there are unknowns in regard to how the earth will respond to increased CO2 levels. Sulphates, aerosoles, clouds, forest cover and ice albedo all are understood within boundaries and the paleoclimatic record indicates that 1.4 to 4.5 (ie a mean of 3) is the range of temps available for a pre industrial doubling of CO2.
is climate is more sensitive than reported (IPCC often criticised for being too conservative) then we might experience 0.3/0.4 C per decade warming as oppossed to the current 0.2C and if levels get higher than yer 1.6C is already guaranteed (hence adapting to CC is needed) and another 0.4c is not far away. only one or two decades depending on sensitivity.
Edward Greisch posts:
[[ 8 years is a very tough timetable to stop the building of coal fired power plants and replace some coal fired power plants with nuclear.]]
You’re right. We’ll have to replace them with conservation, solar thermal, photovoltaic, wind, geothermal and biomass.
“it’s important not to make judgements based on an individual paper, but instead to watch for the follow-up”
Agreed. New science needs some time to season (i.e. weather criticism) before it can be safely taken up by policy makers. Haste makes waste.
[Response: To amplify on that sentiment, though, I do think it is fair to take into account the track record of the investigator, the degree of support for the idea that can be seen in prior work, and the nature of the argument when making a jugement of what the long term assessment is likely to be — while waiting for that assessment to materialize. Thus, while one ought to wait to see how the results of Camp and Tung hold up, given that Tung is quite meticulous and given the straightforward nature of his analysis of the solar cycle in the AR4 archive, I feel pretty confident that that result is likely to hold up. We shall see. Whether the observed solar cycle in surface temperature is as large as .17K (as in Camp and Tung) or more like .1K (many previous estimates) is somewhat more in doubt, as is their interpretation in terms of low thermal inertia and high climate sensitivity in energy balance models. –raypierre]
Some press coverage of this affair appeared in Liberation today. Check out http://www.liberation.fr/transversales/futur/298894.FR.php
And a report also in Le Monde:
http://www.lemonde.fr/web/article/0,1-0,36-991405,0.html
I think maybe the word “fraud” used in the headline of this article is going further than can actually be demonstrated at this point, which is why I preferred the judicious phrase used by Damon and Laut in connection with other work of this type: “A pattern of suspicious errors.” Other than that, I think the article captures the nature of the issues pretty well.
Raypierre, I agree that fraud is too strong. I might choose delusion. The really odd thing is that with 10% of the effort the authors have devoted to deceiving themselves, they could actually understand the physics. It really does emphasize the importance of consensus in science.
Edward Greisch wrote: “8 years is a very tough timetable to stop the building of coal fired power plants and replace some coal fired power plants with nuclear”
Barton Paul Levenson: “You’re right. We’ll have to replace them with conservation, solar thermal, photovoltaic, wind, geothermal and biomass.”
Conservation and efficiency are by far the fastest and most cost-effective ways to reduce carbon emissions, especially in the USA where we are profligate wasters of energy, and thus can easily reduce our energy consumption and associated emissions by applying readily available conservation and efficiency measures.
On the supply side, there is some very good news today: the California-based company Nanosolar has begun shipping their thin-film photovoltaic panels. The first panels will be used in a solar power plant in Germany, which will initially produce 1 megawatt of electricity. According to Nanosolar, their new ultra-cheap, flexible thin-film PV panels will enable the production of solar electricity for less than one dollar per watt — less than the cost of fossil-fuel or nuclear generated electricity. While the initial product offering from Nanosolar is for utility-scale applications, the same technology will eventually be available for small-scale distributed applications, including residential rooftops. Nanosolar says that their entire planned production of thin-film PV panels for 2008 has already been sold.
And yet more press:
http://www.20minutes.fr/article/202189/Sciences-Les-sceptiques-francais-du-climat-accuses-de-tricherie.php
#21, 22, 25…
Yes, a good example of what we call “pensée unique” in France: tomorrow, it will be l’AFP, Le Figaro, L’Humanité, this week-end Le Nouvel Obs, L’Express and Le Point, next month Science & Vie, La Recherche and Sciences & Avenir… But I’m pleased to sea how my French colleagues are now vigilant for dubious curves and how they’re interested in solar effects on climate. I’m sure that beyond the polemic, they will from now on inform their readers on real climatic debates in a more balanced way.
raypierre (#25) wrote:
Swweeet!!
I had to run it through Google Translator, I’m afraid. (My apologies, but I took German due to Kant — and didn’t do terribly well at that. Not that much point, I suppose: I am told that Kant gets taught in the English translation in Berlin — as he is more easily understood by native German speakers that way.)
While I am sure that more than little was lost in translation, it looked devastating, hitting a number of the major points — and your critique on Real Climate received a fair amount of space. They did far better job than I have come to expect from the US press — but then it would be quite a different world as of recent if the US press generally rose to the level of mere mediocre performance.
Incidentally, I enjoyed your critique. Quite thorough. And it gives us a bit more insight into the kind of politics which can get that sort of paper in print.
[Response: Without divulging any privileged information, I can say — as one who has himself had some experience as an editor — that another factor in such papers getting into print is that it is often hard to find qualified reviewers who are willing to take the time to delve into a paper like Courtillot’s. Everybody has a lot of demands on their time, and it is much more productive to ones’ research to spend time with good and interesting papers rather than bad ones, especially in view of the fact that there are so many of the latter. For every four papers like this that get stopped in review, one is bound to get by, just through exhaustion of the reviewer pool. Some editors are better than others at tapping into a pool of suitable reviewers, and it helps if a paper is bad in an interesting way, and is laid out so one can at least follow the argument. No reviewer is likely to go through the amount of work Bard and Delaygue had to do to dig out the flaws in Courtillot’s paper, especially since work of that sort done at the review stage never sees the light of day. –raypierre]
I have been in communication with Rob van der Hilst, who is understandably more than a little unhappy about my commentary on the way he handled the Comment/Response to Courtillot et al. I somewhat regret bringing up the issue of possible problems with the EPSL editorial process at all, given that it is a side-show to the main issue at hand, which is exposing the nature of the arguments in Courtillot et al. However, I thought it necessary for the readers to know the full history of how the true data source for the supposed “Tglobe” came to be known, and of who first pointed out the problem with Courtillot’s use of the Tobiska solar data series. I did not see any way to explain this without bringing Rob’s name into it, given that he was the responsible editor and made all the decisions. I tried to present the sequence of events as best I could infer them from what appeared on the Elsevier web site, augmented by a few queries I put out (none of which went directly to Rob; remember I am a scientist, and not schooled in the reflexes every professional journalist has).
I invited Rob to give his view of the editorial process here, but he did not want to pursue it in this forum. In lieu of that, here is a second try by me, to present the basic facts of the matter in the most neutral possible manner, taking into account Rob’s perception of what went on. I would be interested in what the readers make of this. In the same spirit, I emphasize that I stated Rob’s professional connection with IPGP purely as a matter of relevant background information. I of course am not a mind reader, and can have no knowledge of what role this connection may or may not have played in the way Rob chose to handle this. So here goes:
As I stated originally, at the proof stage of their Comment, Bard and Delaygue had the opportunity to see Courtillot’s Response. Noting the peculiarities in the Response regarding the solar and Tglobe dataset, they felt that in light of this new information an additional remark was called for, and for want of any other vehicle proposed handling it as a “Note added in Proof.” This is quite standard editorial practice when there is a late-breaking development, and when such a note is added to a regular journal proof, the Editor generally has leeway to accept or deny the note. In this case, the Editor articulated the opinion that the Comment/Response cycle was closed and no further information should be introduced, and the Editor feels he exercised his Editorial discretion to deny the approval of the Note Added in Proof. Bard was left with a very different impression of the outcome, particularly in view of the fact that his Comment was posted on the Elsevier preprint section with the Note intact. As far as I can tell, there was no further communication with Bard and Delaygue concerning the matter until the final prepublication proofs were sent. When it came time for final publication, the Note had disappeared. I suppose Rob would say that he intended this all along, and that the appearance of the Note on the posted Preprint version was an oversight; I am only guessing about that, though. At some time before final publication, van der Hilst, having declined to exercise his editorial discretion to re-open the “closed” Comment/Response cycle when it came to Bard and Delaygue’s “Note Added in Proof,” decided to exercise his editorial discretion to the extent of re-opening the “closed” cycle in order to allow Courtillot to modify his Response so as to take into account the points raised by the Note Added in Proof.
So there they are, just the facts, Ma’am (as they used to say on Dragnet), in the most unvarnished way I know how to present them, and as best they can be known to me. I would be very interested in what conclusions the readers draw from this sequence of events, what may or may not have been lost by handling things this way, and how the readers feel the matter should have been handled, if they do not feel it was handled correctly.
Thanks for the great article. Indeed, I do find it very suspicious that you see the same names over and over, regurgetating the same nonsense over and over. Same stuff as well- water vapor, CO2 lag, 1940-70 cooling, MWP, ad nauseum. The number for S&W solar number keeps getting higher and higher, I’m sure we’ll see a “70% solar contribution” in some 2010 paper, but who knows
One point that has been made on and off before, but needs to be repeated was the point on the concept that there is only one way to look at radiative forcings: the sum of them all. The scientific basis for anthropogenic climate change is there, but now we find some unknown forcing, or some forcing is a bit higher than we suspect, etc. You can’t just now “replace CO2.” The logic of “the sun is going up, so now we can have a picnic because there will be no problem or we can’t do anything about it, and CO2 is falsified” is high level nonsense: I’d be much more worried if the sun is going up, and CO2 is going up. Heck, if Svensmark or Shaviv is right, Scafetta and West is right, the CO2 forcings are right, Courtillot is right, etc then we really are screwed!
I do need clarification on one point though. Say we have Total forcing represented by Ft which equals 10 W/m-2 (hereafter just the number), where F1 + F2 + F3 = Ft. Say we think F1 = 4, F2 = 11, and F3 = -5 (so 10). Now say that it turns out through further tests that F1 really equals 7, not 4. First of all, don’t we know the “total” forcing, (that is we can quantify the radiative imbalance and tell you how much Ft is, like now it is ~1.6 W/m-2). So that means if a partial forcing goes up, doesn’t another partial forcing need to go down (or throw in a negative F4)? Since we know Ft = 10 ( i.e. back to the real world, don’t we know the imbalance is 1.6 W/m-2 now so dont all forcings need to add up to the imbalance?)
[Response: We have estimates of the imbalance through looking at heat storage in the ocean and looking at the surface energy budget, but I’ll defer to Gavin on how accurate those estimates are. He was directly involved in that area. So far as I am aware, satellite data isn’t yet accurate enough to detect the top-of-atmosphere imbalance, but somebody please correct me if there are new developments in this area. But the more important thing to note is that the imbalance is emphatically not the same as the radiative forcing. If you doubled CO2 and let the system come into equilibrium, the imbalance you’d measure from space would be zero — but there would still be about 4 W/m**2 of radiative forcing from the change in CO2. Radiative forcing is not a zero-sum game in the sense you laid it out, but it is fairly standard practice to use the observed temperature record to put some constraints on both climate sensitivity and the magnitude of unknown forcing. For example, we know the past CO2 radiative forcing to very high accuracy, but there are more uncertainties in the aerosol forcing; applying a consistent climate sensitivity to both CO2 and aerosols, you can get a match to the observed record for a range of different supposed aerosol forcings, but you can’t take it too far. If you set the aerosol forcing to zero you don’t get the mid-century interruption of warming, and if the aerosol forcing were allowed to get as big as, say, 10 W/m**2 you would get excessive cooling unless you imposed a very low climate sensitivity — which would then make it impossible to reproduce the post-1970’s warming. You get the picture. Crowley’s paper is a nice one to read to see how this is done, since it’s done with an energy balance model where you can see the big picture without a lot of complications –raypierre]
I’ve had enough experience with the referee process not to believe for a nanosecond that the reason this paper was published was anything other than the name of the first author, pure and simple. I’m sorry, RayPierre, but I don’t buy your argument. Editors are politicians, like it or not.
#29 response- Thanks, I’ll look into it
#30
things that are wrong, and things that seemingly have the intent to deceive makes their way into peer-review quite a bit (moreso the former, which is good). I am a bit disappointed in the peer-review process, actually, given the abundance of literature that RC and other places need to cover because of errors. I also don’t think that the name of the guy was a significant part of Ray’s argument, and you can focus on issues like the objection paper, as well as why Damon and Laut had to write a paper on “suspicious” papers.
Raypierre,
Whatever measures you adopt for tactical reasons, please never abandon the 10 foot baguette. ;)
Raypierre, I have also had a little recent experience with a bad paper that “slips through the cracks”. A paper I reviewed had some very serious errors. I pointed these out and suggested a major revision, thinking (indeed hoping) that the paper would go away. Alas it did not, and the second draft was much improved, though still flawed. Again, I pointed out the flaws and made suggestions. Like a bad penny, it came back for a third revision, then a fourth. After several cycles of this, with the journal editor under significant pressure to publish from the authors and under deadline, I gave my final review, detailing the remaining errors, suggesting corrections (some quite involved), but not requiring further review. I was conflicted. I know the final paper will be quite flawed. It will present some interesting ideas from a different perspective than the community gets, though, and the final product will be much better than it started out. Perhaps I should have held firm against publication of a crappy paper, but one must make a judgement whether the modicum of good in a paper might outweigh the bad. I decided that my peers are big boys and girls and will be able to sort out the good from the bad.
Perhaps Rob made a similar judgement, and I’m certain he is chagrined to find this playing out on the pages of the French media. Unfortunately, however, I think he was in a no-win situation. Had he rejected the paper or had he allowed the Note, he would have been accused by denialists of “being part of the conspiracy,” and that would have played out on the pages of the National Post, the WSJ and other contrarian rags (remember Benny Peiser’s ill-fated “rebuttal” to Oreskes’s work). I suspect that editors face considerable pressure to allow into print the work of contrarians unless it is beyond all redemption. So crap gets into print precisely because contrarians demand a voice and all there is to publish, once you remove all the real climate science, is crap.
Hello, and thanks Ray for this interesting saga on aour french climate-skepticals. I just updated my website to signal that two major daily newspaper in France have published article yesterday on this story. Here are the links for french speaking people…
On Liberation, http://www.liberation.fr/transversales/futur/298894.FR.php
On Le Monde, http://www.lemonde.fr/web/article/0,1-0@2-3244,36-991405@51-853716,0.html
#11 and “Civilization may fall anyway well before 2050, but we can avoid going extinct by 2100.” I read SIX DEGREES last summer, and from what I recall, Lynas did not imply this. His idea (as I remember it) was that we would likely COMMIT the world to great harm (as 55 or 251 mya) if we reached 3C warming, in that slow positive feedbacks would eventually push us to 4C, then 5C & 6C. However the harms would take a long time (centuries, perhaps millennia) to completely play out, even after reaching 6C. More like bleeding to death from a thousand small wounds than a sudden (within this century) total wipe-out.
And I think Ray is right, that no matter how bad it gets, we could still lessen the harm, or slow its realization.
raypierre says:
“At some time before final publication, van der Hilst, having declined to exercise his editorial discretion to re-open the “closed” Comment/Response cycle when it came to Bard and Delaygue’s “Note Added in Proof,” decided to exercise his editorial discretion to the extent of re-opening the “closed” cycle in order to allow Courtillot to modify his Response so as to take into account the points raised by the Note Added in Proof.”
this is a clear-cut case of editorial misconduct. It has happened to me too
in this particular way with the addition that after the comment/response
cycle was closed our refereed reply was shown to the commenter who
was then allowed to modify the comment which was subsequently (re-)accepted
with no further review. At that point we threatened to sue the editorial
board (the comment touched on being libellous as we were advised by an
in-house lawyer). After the threat we got a letter from the editor that
stated the commenter had withdrawn the comment after considering our
reply. We let the matter end there, although we have the editor
admitting on letter-head to a violation of the editorial policy that governs
commets/replies. Go figure.
>>>>
Courtillot also cites an atmosphere-ocean model simulation by Zorita et al. (2004) as support for his claim that models fail to represent the 20th century response to solar or magnetic variability. However, as discussed here and in the peer-reviewed references cited therein, this simulation suffers from an inappropriate initialization which leads to a spurious cooling in parts of the run, and a large climate drift requiring detrending of the output before analysis. Besides that, the model explicitly neglects anthopogenic aerosol forcing, so how could one expect it to get 20th century climate right?
>>>>>
Dear Raypierre,
your and Courtillot’s discussion of the Zorita et al (2004) simulation are inaccurate. Contrary to Courtillot et al (2007) the NH T trend simulated by the model in the 20th century does by no means differ from observations by a “factor of three”, but rather is about 10% larger than in the observations. I do not know on which grounds Courtillot et al (2007) reach such conclusions.
Your are right when you state that that particular simulation was not driven by anthropogenic aerosols and therefore a match to observations should not be necessarily expected. However, the simulation does not show a “spurious drift” and the output does not need to be detrended at all. You have probably mixed up several things here: another, different and longer, simulation with the same model – which actually your link refers to; the issue of detrending or not detrending in the context of climate reconstructions; and the issue of the use of flux-adjustment (the output of some models that do not use it may display a spurious climate drift – our model is however flux-adjusted).
Unfortunately the Courtillot et al paper and this part of your post can be quite confusing to the readers in this respect
[Response: Thanks for the clarification. I read your paper (which had some very interesting points, by the way) and from the description there it looked to me like the simulation was the same one as we discussed previously on RC. You’re the one who knows for sure, though, so I’ll defer to you on that. Given the number of papers I needed to read in order to check the claims made in Courtillot et al., I was not able to spend as much time with each one as I would have liked. I will point out that with regard to the claims made by Courtillot regarding your results, flux adjustment introduces similar issues to detrending. These issues are not necessarily important for the things you were trying to do with the model, but they are relevant to the way Courtillot was trying to spin the results. –raypierre]
Gavin, Ray, anybody,
I’ve written a popular article on the saturation argument. It’s got some math in it, but nothing harder than algebra. About 2400 words. Can you think of any place that would take something like this? I emailed Physics Today, but they never replied.
[Response: You might try American Scientist, which is the house magazine of the scientific honor society, Sigma Xi. –raypierre]
Raypierre and also Chris (comment 29) noted that “sceptics” should not just magically forget about or omit the established greenhouse gas physics when trumpeting exotic solar explanations for current warming. But as far as I can tell, most sceptics don’t flat out deny greenhouse gas warming, but they incorporate their “extra” forcing by assuming a lower climate sensitivity. That way they don’t have to completely deny basic physics, and they throw the exotic stuff in by using the relative uncertainty of climate sensitivity as a free parameter so to speak. By doing so they may actually deny the constraints that there are on sensitivity, that I’m not sure about.
FWIW, journal editors have the role of bringing the maximum information to the readers not acting as judges between the parties with the sole and vital exception that they must make an editorial decision that the the correction is worthy of consideration. It is all to easy in a controversy to try and cut it off (which is the mistake that the editor here made), but I see nothing wrong with showing the information to the original authors for their comments. There is no “deadline to meet” and the process takes as long as it takes. That does not mean that when one of the parties starts sprouting nonsense the editor cannot rule the nonsense offsides, or cut off the process when no new information is being added.
I have been on both sides of this, and consider myself fortunate in the editors that handled the situations. I have, through the guidance of those editors, learned a lot about these situations. A word of advice to those submitting comments – keep it short, only comment on things you definitely can prove.
[edit]
#24, SecularAnimist: Here’s another company in the thin film solar panel business that’s just getting started. They expect to be selling their panels for about a dollar a watt also.
http://www.toledoblade.com/apps/pbcs.dll/article?AID=/20071219/BUSINESS03/712190411
I think it’s pretty interesting that just today, Pres. Bush said that nuclear is the “best” way to combat global warming. Oh well.
My Repulican friends couldn’t wait to inform me of the Senate report just posted to Imhof’s Press blog:
http://epw.senate.gov/public/index.cfm?FuseAction=Minority.Blogs
If anyone ever needed a complete guide to the “scientific Denialist’s”, this report should do it.
The authors of the paper under discussion are prominently featured.
[Response: Yeah, this is pretty much a non-story. All you have to do is look at the names and some of the creative accounting used to bulk up the list. For example, they’ve added pretty much all of the names of Lomborg’s “Copenhagen consensus,” even though most arent’ scientists. Then, there’s the trick of including, say, all the co-authors of the Courtillot article, even though some of them (like Fluteau) certainly would not describe themselves as skeptics. There is Avery, with his “unstoppable global warming” nonsense (and no peer-reviewed scientific publications). There are a few real scientists, like Allegre and Courtillot (’nuff said about that), plus a lot of nonentities whose lack of credentials you can spot just by running their names through Web of Science. If this is the best 400 names that Inhofe can scrape together, then the denialist crowd is really in bad shape. I don’t think it needs any more comment than that. –raypierre]
Raypierre, re. 29 “We have estimates of the imbalance (energy) …looking at the surface energy budget,”
For determining surface energy budget, do you mean by measuring borehole evidence or other means?
http://www.sciencemag.org/cgi/content/full/297/5579/206?ck=nck&ijkey=TgKv6oKI2SgcE&keytype=ref&siteid=sci
[Response: No, for long term heat storage it’s only the sea surface that counts. What I had in mind were estimates of net surface imbalance using boundary layer models driven by meteorological observations. –raypierre]
RE #42 Paul Middents:
Yeah that tidbit was brought to my attention as well. It is on today’s Drudge Report. Everything in one overwhelming re-package. BTW, the paper under question is exactly the kind of item that will be brought up again and again to show that there is still a scientific controversy. Excellent critique, Raypierre, and thanks for all the dispatches from San Fran.
A bit off topic for this thread, but certainly relevant to RC:
Here’s an clever approach to getting people to think about AGW and its consequences:
http://www.switched.com/2007/12/19/high-school-teachers-global-warming-video-a-youtube-hit/
Of course, he probably won’t change the minds of many die-hard skeptics, as is evident from the comments listed below the video window.
Editorial Handling of Comments and Replies
However necessary, a Comment and Reply procedure always demonstrates a failure of the publication system: noble material appears in full-blown regular papers. During 11 years of editorial tenure in major geophysics journals, I always agonized about how Comments and Replies should be handled, especially when one of the parties is as prejudiced as in the present case. I stick to two principles: (1) the Comment and Reply exchange thread can’t go on forever, and (2) the defense (Replyer) should have the last word and not be opposed arguments that they can’t respond to. Clearly these principles are a potential source of frustration for the Commenter when the Replyer’s good faith is an issue. Finding editors of high stature is already a very difficul task but if you now hold them liable for all the mistakes, lapses, and twists slipping into each manuscript, this task will soon prove impossible.
[Response: Very true words,and indeed an Editor is in a bind when one of the parties (the Reply party, in this case) is acting in bad faith. If I have some complaint about the way the Editor handled this, it is that to me the way out of the dilemma was much more indulgent to the bad-actor (Courtillot) than to the white-knight who put in the work to dig out the errors. Further, it was surely relevant information that in the initial Reply, Courtillot attempted to deceive yet again, by giving the name of a purportedly global dataset that turned out not to exist, and was certainly not the one used by the authors. It was only on the second attempt that the name of the real data (which turned out to be neither global nor annual) came out. Surely this history is relevant information, which should have been preserved. In fact, it was preserved, but only because, by luck, I happened to be writing my post based on the earlier version of Bard and Delaygue’s comment, before the Note was deleted and Courtillot’s Reply was modified. In the end, some critical information did get out, but if it werent for these fortuitous circumstances, some critical information would have been lost as well. –raypierre]
I have no envy for editors, and I am sure it takes herculean strength to resist the arm-twisting they’re subjected to, but time and again one sees that
(1) Progressive, unconventional ideas from young researchers are biased against,
(2) Regressive, outdated ideas from old farts who should just get out of the way, are biased for.
In other words, scientists are actually phenomenally conservative; something the right wing clearly just doesn’t get.
You can still find papers being published from time to time on non-cosmological redshifts of quasars, for example, which is total nonsense, but it is a regressive idea held onto by old cranks who don’t believe in general relativity and have names big enough to get their ideas to press.
The whole peer-review system is a mess. The public thinks that being peer-reviewed is a stamp of approval, which it isn’t, and editors themselves aren’t clear on precisely what refereeing/editing is supposed to accomplish. Last time I looked, the major astrophysics journals did not publish any basic position statements on what, exactly, being approved for publication was supposed to significate. So how is the public supposed to figure it out if the scientists themselves don’t even know? I’d love to hear it if the situation were different in geophysics.
[Response: I wouldn’t go so far as to say that it’s a mess, but I do think that it is overloaded. The journal Atmospheric Chemistry and Physics has an interesting new angle on this. Reviews are anonymous, but they are posted for the public. In addition, the general readership can post comments and suggestions on a preprint, just as on a blog. As I see it, though, the main problem is that people are writing too many papers, many of which are just not that significant. Everything would work more smoothly if more people held their fire until they had something fairly important to say. Unfortunately, a lot of the reward structure is not set up to foster that. –raypierre]
Dear Raypierre,
Thank you for your post and clarification. I completely agree with what you wrote (and also with Bard & Delaygue’s paper…).
Just to add my grain of salt, I recently (i.e. before your posts) had a conversation with somebody at IPGP who explained me the context of Courtillot et al.’s EPSL paper:
In 2003, Gallet and others (incl. Courtillot) published in EPSL a study on paleomagnetism from archeologic data over the last 3 millennia. Le Mouel, Kossobokov and Courtillot published a paper in EPSL in 2005 on the magnetic field and climate. For some reason Courtillot wanted to paste the two papers together, hence the list of authors. The latter 2005 paper made (admittedly too little) noise in the climate science community. But nobody, including Edouard Bard, seem to have had the time to respond to the paper. I just remember that I was extremely perturbed by the abundant use of the adjective “significant” in the text, while no real quantification of this signifiance was provided. The other thing that should strike any non specialist is that temperature (whatever its true reference!) always leads the solar and geomagnetic data by a few years… oh well, I just told to myself that the paper would not take long to be forgotten…
So, it appears that the present controversy originates from the 2005
EPSL paper. In the unlikely event that the 2007 paper is withdrawn or whatever, there remains the 2005 paper, which everybody seem to have forgotten and contains the very same flaws outlined by Raypierre, and Bard and Delaygue in the 2007 paper. Incidentally, this also says that Courtillot never cared to improve his physics of climate between 2005 and 2007.
To conclude, with colleagues, we had thought of nominating Allègre and Courtillot for an Ignobel award (say, in the climate nonscience
category), but it turns that the Ignobel awards are given to true
results (albeit strange!) and they probably would not be qualified.
Best to all and joyeuses fêtes,
[Response: The remark about the 2005 paper raises another troubling question about the 2007 Courtillot et al paper — As far as I can tell, there is little if anything new in this paper beyond what has been published by the authors in the authors’ previous works. In essence it is a review article, not a research article. Yet, it does not appear to be one of the invited EPSL “Frontiers” reviews. Does EPSL now publish unsolicited submitted review articles? I wasn’t aware that it did. So how did it happen that this paper was even allowed to enter the review process, let alone survive it? Curioser and curioser. (et joyeuses fêtes!) –raypierre]
To #47
To paraphrase Churchill, peer review is the worst form of editorial handling except all the others that have been tried. We can also joke about either old farts resisting change (fair enough) or that younger ones see themselselves as creative because they never read anything more than 5 years old (equally true). Junior and senior scientits having a different perception of science is a good thing after all: energy arises from differences in potential. The science you inherit today was left to you by this less than perfect system. We all fell victim to abusive or incompetent reviewers. Good reviewers (and editors) are rare and overtaxed. I fear, however, that if they know their reports will posted, many of them simply will decline the review. Don’t throw out the baby with the bath water!
Francis Albarede, A very thoughtful post. The thing is that peer review typically serves many functions. Yes, it should serve as a barrier preventing very bad science from seeing the light of day, but more commonly its role is to take decent science and push it into the realm of good science. Often, peer review is problematic for either really great or really bad papers. In the former case, the author may be saddled with lots of trivial changes, or worse yet be asked to make changes that weaken the paper. And in the case of very bad papers, the sheer persistence of an author can result in the paper tunneling through even if it is deeply flawed. Fortunately, science is strong enough to deal with these failings–enough scientists will recognize great work once it is published to ensure it is incorporated, and crappy papers usually get the obscurity they richly deserve (it’s a lot harder to tunnel your way to general acceptance than it is to tunnel through the review process).
The main problem comes when you bring in the lay public or reporters, who often are not sophisticated enough to separate the wheat from the chaff and who often uncritically accept peer review as a seal of approval by the scientific community. People need to understand–peer review is a minimum threshold. The real stamp of approval is scientific consensus.