ABSTRACT: Completing the loop: from data to decisions and back to data.

CGRG Bibliography of Canadian Geomorphology
Search Results

Author :

Date :

Title :

Publication :

Issue :

Page(s) :

Abstract

The role of science in decision-making has been compared to the role of meat in a hamburger. By itself, it is messy, but when contained by the bun of policy it becomes more palatable. Extending this metaphor, data can be thought of as the cow that is processed into meat by the scientific process. However, the topic of where the meat comes from rarely comes up in polite conversation. Unfortunately, there is a ‘tragedy of the commons’ occurring— everyone wants a piece of the cow but no one wants to feed it. Furthermore, no one wants to buy a bull that will allow the cow to reproduce. The cow is languishing for the very reason that grain given to the cow is at the expense of the starving graduate students needed to produce prodigious volumes of output. In the meantime, information husbandry is left in the hands of bureaucrats who are accountable only for their ability to balance a budget and who are not held accountable for the legacy they leave.There is incessant demand for quick answers to complex questions. This appetite for ‘fast food’ has resulted in resources being diverted from information husbandry to computer modeling (Hartemink et al., 2001). Scientists have to resort to data scavenging—obtaining whatever ‘road-kill’ they can find to grind up into meat to serve policy objectives. Contemporary environmental science is calorie-rich but low in essential nutrients. When more money is invested in science it results in a ‘super-size’ meal that has extra bulk provided by modeling, but with little added nutritional value. This is because the scientific community is focused on their contemporary needs—to fund graduate students and to publish papers, forgetting the ethics of their profession, which would have them leave a rich legacy upon which their prot´eg´es can build their careers. We are consuming the information legacies of previous generations, but leaving the soil barren for future generations. Well-maintained data appreciate in value like a vintage car. In contrast, model output is like an ice cream cone on a hot summer day. It is intended for immediate consumption with no residual value. Many modelers view the world through the lens of their model algorithms, and sometimes this view of reality is as if seen through a kaleidoscope. In the absence of independent observations of reality how is any modeler to know whether their view of reality is realistic? As our models increase in sophistication we should be investing in more comprehensive monitoring to shed light on how well the models reflect reality. Industrial-scale data production is replacing gauge data as our primary source of information about reality. We can provide cheaper ‘fastfood’ if we are supplied by these data factories, but the sheer volume of remotely sensed data is overwhelming our ability to understand those data. The algorithms that convert any remotely sensed electromagnetic signal into a measure of some hydrologic variable are useful but they are also imperfect. We now understand many of the problems complicating the simple conversion of a radar signal into an absolute quantity of precipitation but only after decades of concurrent gauging since the notion of radar-derived precipitation measurement was first conceived (?S ´alek et al., 2004; Saltikoff et al., 2004). If we do not monitor for the infection of technological artifact in remotely sensed estimates of evapotranspiration, soil moisture and snow water equivalent, how many ‘mad-cow’ diseases will work their way, undetected, up the food chain to eventually lay waste to hydrological science? I fear the 2000s will be remembered as the last decade of real hydrology. We entered this decadewith hydrology based on data but we will be leaving this decade with pseudo-hydrology based on pseudodata. As reliable data supplies dry up, scientists are resorting to sophisticated data assimilation models (e.g. Rodell et al., 2004; Fekete et al., 2004)—models that produce numbers consistent with what we think the data would show if we actually had data. These products are frequently called Gridded Model Output (GMO), though a sharp-witted colleague refers to GMO as being ‘genetically modified observations’. With each layer of ‘Franken-data’ added in a simulation, we become further removed from reality. At the rate we are going, by the 2020s we will use model output as inputs to drive all of our models, and we will verify our models with output from other models, eliminating the requirement for any data at all. We are on a journey from the world of modern hydrology, which is evidence-based, with deep roots in classical philosophy and ethics, to a world of post-modern hydrology, which is model-based with no ethical foundation that has ever been articulated. Whereas modern hydrology is governed by the laws of nature, post-modern hydrology is governed by Moore’s Law as we put more faith in the ever increasing processing speed of our computers (Moore, 1965) than in our ever diminishing field observations (e.g. Shiklomanov et al., 2002; Fekete and V¨or¨osmarty, 2002; V¨or¨osmarty et al., 2002; Lanfear and Hirsch, 1999). There is little assurance of a safe landing in the world of postmodern hydrology, and there is no clearly identified fail-safe point from which we can be sure to return safely to modern hydrology with our available data reserves, if needed. There may be hope. Each and every hydrologist needs to imagine what their information needs might be in 20 years time, and from that perspective consider what monitoring decisions need to be made today in order to supply those information needs (e.g. Kirchner, 2006). Having evaluated the information requirements of future hydrologists, strategic decisions can be made about what data need to be collected, where, and at what time- and space-scales. Data are too important to be left at the mercy of bureaucrats. Scientific data leadership has to come from the whole community, and decision-makers must be informed by hydrologists on how to make the supplyof relevant data sustainable. Strangely enough, if we can devise a strategy to provide future generations with the data they will need, those same data will serve our contemporary needs admirably.

Bibliography of Canadian Geomorphology