Migration towards Bangladesh coastlines projected to increase with sea-level rise through 2100.

To date, projections of human migration induced by sea-level change (SLC) largely suggest large-scale displacement away from vulnerable coastlines. However, results from our model of Bangladesh suggest counterintuitively that people will continue to migrate toward the vulnerable coastline irrespective of the flooding amplified by future SLC under all emissions scenarios until the end of this century. We developed an empirically calibrated agent-based model of household migration decision-making that captures the multi-faceted push, pull and mooring influences on migration at a household scale. We then exposed ~4800 000 simulated migrants to 871 scenarios of projected 21st-century coastal flooding under future emissions pathways. Our model does not predict flooding impacts great enough to drive populations away from coastlines in any of the scenarios. One reason is that while flooding does accelerate a transition from agricultural to non-agricultural income opportunities, livelihood alternatives are most abundant in coastal cities. At the same time, some coastal populations are unable to migrate, as flood losses accumulate and reduce the set of livelihood alternatives (so-called 'trapped' populations). However, even when we increased access to credit, a commonly-proposed policy lever for incentivizing migration in the face of climate risk, we found that the number of immobile agents actually rose. These findings imply that instead of a straightforward relationship between displacement and migration, projections need to consider the multiple constraints on, and preferences for, mobility. Our model demonstrates that decision-makers seeking to affect migration outcomes around SLC would do well to consider individual-level adaptive behaviors and motivations that evolve through time, as well as the potential for unintended behavioral responses.

Altered osmotic thresholds for vasopressin secretion and thirst in human pregnancy.

Osmoregulation was studied in eight women during late pregnancy and again 8-10 wk postpartum. Base-line plasma osmolality (Posmol) was significantly lower during (280.9 +/- 2.1 mosmol/kg, SD) than after (289.4 +/- 2.1 mosmol/kg) pregnancy yet 24-h urinary volume and plasma arginine vasopressin (PAVP) measured in vasopressinase-inactivated blood was similar in both groups (pregnancy, 1.39 +/- 0.56 pg/ml; postpartum, 1.25 +/- 0.62 pg/ml). After 12 h of dehydration PAVP rose similarly and significantly both during (2.25 +/- 0.81 pg/ml) and after (2.89 +/- 1.19 pg/ml) gestation, and Uosmol was similar on both occasions (pregnancy, 779 +/- 121 mosmol/kg; postpartum, 784 +/- 102 mosmol/kg). When Posmol was increased by the slow infusion of 5% saline PAVP increased as soon as body tonicity did both during and after pregnancy. PAVP correlated significantly with Posmol in each subject (range of r, 0.75-0.99) and the mean regression lines [pregnancy, PAVP = 0.32 (Posmol; -279), r = 0.79; postpartum, PAVP = 0.38 (Posmol, -285), r = 0.86] demonstrated that the apparent osmotic threshold for AVP secretion was 6 mosmol/kg lower during than after gestation. Similarly the Posmol at which the subject experienced a conscious desire to drink was lower in pregnant (287 +/- 1.6 mosmol/kg) compared with postpartum subjects (298 +/- 2.0 mosmol/kg; P less than 0.001). These data demonstrate decreased osmotic thresholds for AVP release and thirst during human pregnancy and explain why gravidas can maintain their new lower Posmol within narrow limits.

A comparison of gas-liquid chromatography and a titration method for the determination of plasma non-esterified fatty add levels during pregnancy.

The level of plasma non-esterified fatty acids (NEFA) was measured by gas-liquid chromatography (GLC) and a titration method in 194 samples collected during pregnancy and from four days to 24 weeks post partum. Both techniques indicated a similar pattern of changes in plasma NEFA associated with pregnancy. The titration estimates of NEFA level were usually greater than those measured by GLC, and there was some suggestion that the disparity between the methods was increased at the end of pregnancy and was reduced at six weeks after delivery.

Cord plasma glucose and insulin concentrations and maternal-fetal relations.

The concentration of glucose and insulin was estimated in 854 samples of umbilical cord blood plasma and 503 concurrently collected maternal blood samples. The mean cord insulin concentration, excluding all infants born to known diabetic mothers, was 7 muU./ml., but the distribution was skewed with 10 per cent of infants having a value of 12 muU./ml. or more. The giving of intravenous sugar-coating fluids to the mother during labor tended to exaggerate the skewness rather than move the whole distribution to the right. Babies born by vaginal route, whatever the presentation of method of delivery, had slightly lower mean insulin values than those born by Caesarean section. It thus seems unlikely that "stress" is a factor causing high cord insulin values at birth. Other data concerning maternal-fetal glucose and insulin relations are discussed.